Exponentially Weighted Moving Average Control Charts for Monitoring Increases in Poisson Rate

Size: px
Start display at page:

Download "Exponentially Weighted Moving Average Control Charts for Monitoring Increases in Poisson Rate"

Transcription

1 Exponentially Weighted Moving Average Control Charts for Monitoring Increases in Poisson Rate Lianjie SHU 1,, Wei JIANG 2, and Zhang WU 3 EndAName 1 Faculty of Business Administration University of Macau Taipa, Macau * Corresponding Author ( ljshu@umac.mo) 2 Department of Industrial Engineering & Logistics Management The Hong Kong University of Science and Technology Clearwater Bay, Hong Kong 3 Division of Systems and Engineering Management School of Mechanical and Aerospace Engineering Nanyang Technological University, Singapore Abstract The exponentially weighted moving average (EWMA) control chart has been widely studied as a tool for monitoring normal processes due to its simplicity and efficiency. However, relatively little attention has been paid to EWMA charts for monitoring Poisson processes. This paper extends EWMA charts to Poisson processes with emphasis on quick detection of increases in Poisson rate. Both cases with and without normalzing transformation for Poisson data are considered. A Markov chain model is established to analyze and design the proposed chart. The comparison results indicate that the EWMA chart based on normalized data is nearly optimal. Keywords: Poisson Distribution; Average Run Length; Statistical Process Control; Markov Chain 1

2 1 Introduction Besides monitoring continuous data, there are many applications where it is important to monitor a sequence of discrete counts, such as quality control and productivity improvement in manufacturing processes and health surveillance in health care management (Montgomery 2005; Woodall 2006; Tsui et al. 2008). Count data per unit time or area is often described by a Poisson model. For example, the number of nonconforming items in an inspection unit, the number of reported failures per week, and the monthly number of visits to a certain clinic are often assumed to follow a Poisson distribution. The Poisson distribution has only one parameter, rate of count per unit, which represents both the mean and variance of the distribution. A simple charting procedure for monitoring changes of rate of a Poisson distribution is the Shewhart c-chart or u-chart. The c-chart is used when the count data is based on one unit while the u-chart is applicable if it is based on n units (n > 1). Both the Shewhart c- and u-charts make use of only the most recent observations to determine the statistical status of the process. To make use of the past observations, the exponentially weighted moving average (EWMA) and cumulative sum (CUSUM) charts have been discussed. A sample of research demonstrating the use of CUSUM chart for monitoring the mean of a Poisson process includes Brook and Evans (1972), Lucas (1985), White and Keats (1996), White et al. (1997), Rossi et al. (1999), and Chan et al. (2007). In contrast, the work on the EWMA chart for Poisson data is relatively less. Few exceptions include Gan (1990), Martz and Kvam (1996), and Borror et al. (1998). A comprehensive review on attribute control charts was given by Woodall (1997). A control chart is often designed in one-sided or two-sided form, depending on the objective of detecting one directional changes (either upward or downward) or changes in both directions. When there is an increase in Poisson rate, the process is likely to produce more nonconforming items, which in turn incurs heavier loss and higher cost for companies. On the other hand, a decrease in Poisson mean may imply an improvement in the process. Recently, there is a growing interest on detecting increases of disease incidence rate in health surveillance (Tsui et al. 2008). Therefore, in order to prevent the underlying process from producing too many nonconforming items, efficient control charts being able to early signal upward shifts in the Poisson rate would be greatly desired. The one-sided Poisson CUSUM chart, which originates from sequential likelihood ratio test theory, has been widely investigated as it serves the basis for the analysis of overall performance of a two-sided CUSUM chart. However, little attention 2

3 has been paid to one-sided Poisson EWMA chart, while most of the existing literature mainly focuses on two-sided Poisson EWMA charts, see, for example, Gan (1990) and Borror et al. (1998). To fill the gap in part, this paper aims to develop some one-sided Poisson EWMA control charts with primary interests on the detection of increases in the Poisson rate. Unlike a one-sided Poisson CUSUM chart, one-sided EWMA chart may serve as a baseline model to capture the dynamics of process mean levels when observations are counts. Note that the Poisson data is discrete and nonnormal. It has long been an intuitive idea in Statistical Process Control (SPC) to use transformed data for monitoring when the data is nonnormal. For example, Crowder and Hamilton (1992) considered using logarithmic transformation of the sample variance for monitoring changes in variability. Quesenberry (1991a,b,c) considered using the inverse transformation of a cumulative distribution function for monitoring Binomial and Poisson data. The performance of EWMA charts based on the normalizing transformation of Poisson data has not been previously investigated. The question whether the normalizing transformation can lead to better chart performance arises in both theory and practice. Therefore, it is of great interest to investigate the performance of Poisson EWMA control charts in the case when the Poisson data are normalized. The rest of this paper is organized as follows. First, several one-sided Poisson EWMA control charts are developed based on transformations. Next, the performance of various upper Poisson EWMA control charts is analyzed, followed by some design guidelines. An example of male thyroid cancer surveillance in New Mexico is presented to demonstrate the use of the proposed chart. Lastly, some concluding remarks are discussed. The appendices provide the Markov chain model for the upper Poisson EWMA control charts. 2 One-Sided Poisson EWMA Charts Assume that the process observations, {X t, t = 1, 2, }, are independent following a Poisson distribution with mean µ. The process is said to be in control when µ = µ 0 and out of control when the mean changes to some other values, say, µ = µ 1. The traditional Poisson EWMA control chart is defined by Q t = X t + (1 )Q t 1, (1) where is a smoothing constant such that 0 < 1. The starting value is often set to the target, Q 0 = µ 0 while other head start values can be chosen for Q 0 for fast initial response (FIR) 3

4 features (Lucas and Crosier 1982). The asymptotic control limits are often written as follows, i.e., ( ) LCL = max 0, µ 0 L (2 ) µ 0, UCL = µ 0 + L (2 ) µ 0. This chart is two-sided, which can detect both increases and decreases in Poisson rate. For the purpose of quick detection of either upward or downward changes only, the onesided EWMA chart is desirable. The conventional way to modify a two-sided EWMA chart into the one-sided upper/lower form is to reset the EWMA statistic to the target whenever it is less/greater than the target. See, for example, Champ et al. (1991), Crowder and Hamilton (1992) and Gan (1998). Taking the upper EWMA chart as an example, by resetting the EWMA statistic in Equation (1) to µ 0 whenever it is less than µ 0, an upper Poisson EWMA chart can be obtain as E X t = max[µ 0, X t + (1 )E X t 1], (2) where E X 0 = µ 0. We refer this chart as EX chart in this paper. It signals when E X t exceeds the upper control limit h = µ 0 + L 2 µ 0. Note that the Poisson distribution is asymmetric and is not normal. normality, one may consider the linear transformation To achieve better Y t = X t µ 0 µ0. (3) It is often assumed in practice that Y t is asymptotically normal when the in-control mean µ 0 is sufficiently large. However, it is important to note that this linear transformation just changes the location to zero and standardizes the spread of the Poisson distribution to one. It actually does not affect the shape of the Poisson distribution. To obtain a distribution that is close to normal, a nonlinear transformation based on square root is often suggested, see, for example, Anscombe (1947), Freeman and Tukey (1950), and Ryan and Schwertman (1997). The square root transformation M t = X t + c is approximately normal with mean and variance approximated by (Anscombe 1947) µ M = µ + c 1 8 µ, 4

5 and σ 2 M = 1 4 ( c ), 8µ respectively. Note that when c = 3/8, the variance of M t is nearly independent of the Poisson parameter µ. In this case, the variance of M t is stabilized at σm 2 = 0.25, and any change in the Poisson parameter will be transformed into a change in the mean level of M t. For this reason, we use c = 3/8 in this paper while other choices of c values can be similarly analyzed. Assuming that M t is approximately normal, the statistic X t Z t = µ M µ=µ 0 ( = 2 X t + 3 σ M µ=µ0 8 µ ) 8 µ 0 will be close to the standard normal distribution when the process is in control. When X t in Equation (2) is replaced by Y t or Z t, we obtain another upper Poisson EWMA scheme as E Y 0 = 0 (4) E Y t = max[0, Y t + (1 )E Y t 1], t = 1, 2,..., or E Z 0 = 0 (5) E Z t = max[0, Z t + (1 )E Z t 1], t = 1, 2,..., For simplicity, denote the EWMA schemes in Equations (4) and (5) as EY and EZ charts, respectively. The EY/EZ chart signals when EY/EZ exceeds the upper control limit h E = L 2, which is expressed in the same form as the limit for the EX chart. The values of L are generally different from EY and EZ charts. It is easy to see that both EX and EY charts have the same performance since (E X t µ 0 )/ µ 0 = E Y t. For this reason, we will discuss EY chart only but not the EX chart to keep simplicity throughout the remaining of this paper. Recently, Shu et al. (2007) and Shu and Jiang (2008) considered another resetting rule in the one-sided EWMA scheme, particularly for nonnormal data. They suggested to reset the current observation or normalized observation to the target but not the EWMA statistic. This scheme has been shown to have better performance than the direct reset of the EWMA statistic. Take 5

6 the chart based on Z t as an example. The revised procedure first winsorizes Z t and then applies a conventional EWMA scheme to the winsorized data. That is, where Z + t defined as R t = Z + t + (1 )R t 1, (6) = max[0, Z t ] and R 0 = E(Z + t µ = µ 0). A lower-sided EWMA chart can be similarly R t = Z t + (1 )R t 1, where Z t = min[0, Z t ] and R 0 = E(Z t µ = µ 0). Note that if Z t N(0, 1), the mean and variance of the winsorized normal variable, Z + t 1999) E(Z t + ) = 1 2π σ 2 = 1 Z t π. = max[0, Z t ], are given by (Barr and Sherrill Therefore, the asymptotic mean of R t is not equal to zero but 1/ 2π in the in-control situation. To shift the mean of R t to zero, it is convenient to rewrite the EWMA recursion in Equation (6) as and R Z 0 variance as Z + t Z + t with Y + t = 0. Analogously, define Y + t R Z t = (Z + t 1 2π ) + (1 )R Z t 1, (7) = max[0, Y t ], which has approximately the same mean and since Y t follows approximately a standard normal distribution. Now replacing in Equation (7) leads to the control statistic R Y t = (Y + t 1 2π ) + (1 )R Y t 1 (8) with R Y 0 = 0. For simplicity, denote the revised upper EWMA charts in Equations (7) and (8) as RZ and RY charts, respectively. The RZ or RY chart declares an alarm when R Z t exceeds the upper control limit h R = L 2 σ Z + t = L 2 The values of L are generally different from RY and RZ charts. ( ). 2π or R Y t 3 Performance Analysis In this section, we first compare the effect of two types of resetting schemes in the EWMA on the performance of Poisson EWMA charts. Then, the effect of using normalizing transformation on 6

7 the Poisson data is investigated. All the ARL results are obtained using Markov chain approach, details of which are given in Appendices A and B. The discussions will be based on two cases: the case with a relatively small in-control mean of µ 0 = 4 and the case with a relatively large value of µ 0 = 16. The other cases with different values of µ 0 would provide qualitatively the same conclusions, which are not presented here for the sake of simplicity. 3.1 Comparisons between EY and RY charts For illustration, Table 1 compares the zero-state ARL values of EY and RY charts when µ 0 = 4 and 16 and = 0.05 and The zero-state in-control ARL values are maintained approximately 1,000 for both charts. The shift size, δ, is measured in the unit of the in-control standard deviation of the Poisson process, namely, δ = µ µ 0 µ0. From Table 1, it can be observed that for a fixed value of, the RY chart performs closely to the EY chart at small shifts but has much better performance at large shifts. For example, when µ 0 = 16 and = 0.05, the percentage decrease in the out-of-control (OC) ARL of the RY chart at δ = 0.5 relative to the EY chart is 1.52%(( )/32.96) but the corresponding percentage decrease at δ = 3 is 23.97% (( )/3.88). Note that when = 1, the EY chart reduces to E Y t = max[0, Y t ] = Y + t, and the RY chart reduces to R Y t = (Y + t 1 2π ). Clearly, both charts are essentially the same upper-sided Shewhart chart of Y t. Therefore, they would perform the same when = 1. This implies that when increases, the improvement of the RY chart over the EY chart at large shifts tends to diminish. To illustrate this point, Figure 1 further plots the percentage decrease in the OC ARL of the RY chart relative to the EY chart for different values of when µ 0 = 16. From Figure 1, it is clear that the relative OC ARL percentage decrease reduces from about 20%-25% to 2%-5% for detecting shifts of δ 2 as increases from 0.05 to 0.5. Note that the same value of is used for both EY and RY charts in the above comparison, which might not be optimal for either chart at the same mean shift. For this reason, Table 2 7

8 Table 1: Zero-State ARL Comparisons between the EY and RY Charts µ 0 = 4 µ 0 = 16 = 0.05 = 0.15 = 0.05 = 0.15 EY RY EY RY EY RY EY RY δ L= further presents the optimal performance of both charts aimed at detecting shifts of sizes of δ opt = 0.5 and δ opt = 1.5. The optimal values can be obtained from the design guidelines presented in the next section. To provide a more complete comparison, the steady-state ARL is also included. The steady-state ARL refers to the ARL obtained assuming that the control statistic has reached steady state before changes occur, which was computed in a way similar to that of Lucas and Saccucci (1990). Details of the computation are given in Appendix A. From Table 2, it is interesting to note that the optimal values of the RY chart are always smaller than those of the EY chart for detecting the same size of shifts. The RY chart always performs better than the EY chart at the optimized shifts in all cases considered here. In fact, the RY chart performs uniformly better than the EY chart at all shift sizes. The steady-state results provide qualitatively the same conclusion. This observation is clearly an evidence of superior performance by resetting the observation/transformed observation but not the EWMA statistic in detecting increases in the Poisson rate. The observation is consistent with that in the case of monitoring normal process mean (Shu et al. 2007). 8

9 Table 2: Optimal ARL Comparisons between the EY and RY Charts µ 0 = 4 µ 0 = 16 δ opt = 0.5 δ opt = 1.5 δ opt = 0.5 δ opt = 1.5 EY RY EY RY EY RY EY RY = δ L= Zero-State Steady-State

10 Percentage Decrease in OC ARL =0.05 =0.15 =0.25 = δ Figure 1: Percentage Decrease of Out-of-Control ARL of the RY Chart Relative to the EY Chart When µ 0 = The case with normalizing transformation We now compare the run length performance of the upper Poisson EWMA charts based on data with and without normalization transformation. Tables 3 and 4 provide the comparison results for µ 0 = 4 and µ 0 = 16, respectively, with optimally chosen for certain shift sizes. It can be observed that the optimal performance of the upper Poisson EWMA chart based on normalized data is nearly the same as the chart based on un-normalized data. For example, when µ 0 = 4 and δ opt = 0.5, the EY and EZ charts have the respective ARL values of and at δ = 0.5, while the RY and RZ charts have the ARL values of and 31.23, respectively. Clearly, the differences in these optimal ARL values are negligible. This indicates that the normalization transformation of the data merely makes data more normal but not results in any performance improvement of detection. This observation is similar to that commented in Gan and Fan (1995). They showed that the EWMA chart based on the logarithmic transformation of the sample variance is nearly optimal, compared to the EWMA chart based on the sample variance. However, it is surprising to note that the normalization transformation tends to deteriorate the performance for detecting other shifts in most cases. 10

11 Table 3: Zero-State ARL Performance of the Upper Poisson EWMA Chart in Cases with and without Normalizing Transformation When µ 0 = 4 δ opt = 0.5 δ opt = 1.5 EY EZ RY RZ EY EZ RY RZ = δ L= Table 4: Zero-State ARL Performance of the Upper Poisson EWMA Chart in Cases with and without Normalizing Transformation When µ 0 = 16 δ opt = 0.5 δ opt = 1.5 EY EZ RY RZ EY EZ RY RZ = δ L=

12 4 Design of Upper EWMA Poisson Charts In the above comparisons, the RY chart demonstrates some advantages for practical use. It is shown that the RY chart performs better than the EY chart and has performance similar or superior to the RZ chart. To facilitate the design of the RY chart for practitioners, this section provides some design guidelines. The design of the RY chart is a two-degree-of-freedom problem, which involves determination of the values of and L. Similar to the design approach of Lucas and Saccucci (1990), one can select the value of to optimize the detection performance for a specific change in the Poisson mean and then choose L to provide a desired in-control ARL. There is no simple guidelines for finding the optimal parameters of RY charts. Extensive computation was thus performed to numerically search the nearly optimal values of and L for RY charts. Tables 5 to 8 present a list of optimal parameters of RY charts when in-control ARL values are set as 370, 500, 1000, and 1500, respectively. These parameters are adequate for most practical purposes, which are obtained using a numerical search method through the Markov chain approximation. For example, suppose we want to find the optimal parameters of the RY chart for the target shift of size δ opt = 1 with in-control zero-state ARL of 370 and µ 0 = 4. The L value can be paired with a wide range values of = 0.01, 0.02,,(stepsize=0.01) to produce the desired in-control zero-state ARL of 370. The out-of-control ARL at δ = 1 is then evaluated for each of the (, L) combinations. The chart with the combination ( = 0.09, L = 3.044) gives the minimum ARL of 9.0, and thus these parameters are selected to be optimal. The same approach is used to determine the optimal parameters of RY charts for other in-control ARL values and different values of µ 0 and δ opt. From Tables 5-8, it can be seen that small values of are more sensitive to small shifts while large values of are more sensitive to shifts of relatively large sizes, as expected. For example, when the in-control ARL is 370 and µ 0 = 2, from Table 5, the optimal values of increases from = 0.02 to = 0.19 as δ opt increases from δ opt = 0.25 to δ opt = 2.0. Furthermore, it can be observed from Tables 5-8 that the in-control mean value of the Poisson distribution has impact on the choice of optimal parameters for Poisson charts for given target sizes of shifts, especially when µ 0 is small and the target shift size is large. This observation is different from the case of monitoring normal process means. However, this effect tends to be negligible when µ 0 is sufficiently large, say µ From Tables 5-8, it can be seen that the optimal values of at a particular shift are nearly the same for different values of µ For 12

13 example, when the in-control ARL is 370, the optimal choices of aimed at detecting a shift of size δ opt = 1 is always given by = 0.10 for µ 0 16, as shown in Table 5. This is mainly due to the property that the Poisson distribution approaches a normal distribution when µ 0 is large. Meanwhile small values of also make the approximation of the EWMA statistics based on nonnormal data to be close to normal (Borror et al. 1998). 5 An Example of Male Thyroid Cancer Surveillance in New Mexico To illustrate the use of RY chart, the data on the incidence of male thyroid cancer in New Mexico was used here. The data was collected by at the National Cancer Institute and was available through the Surveillance, Epidemiology, and End Results (SEER) Program. This example was also studied in Mei et al. (2010). The annual incidence of male thyroid cancer per 100,000 persons over was presented in Table 9. The main objective of this application is to detect an increase in the incidence of male thyroid cancer. It is well known that the incidence is stable during the period from year 1973 to year 1988 and increases since year For this reason, the first 16 observations during the period were considered to be in control and the remaining observations are out of control. The incidence rate based on the first 16 observations can be estimated as 2 per 100,000 persons. For the sake of simplicity, we assume that the estimate is the true value of µ 0. However, it is important note that estimation error often exists in practice, which would result in negative effects on control chart performance. A detailed review on effects of parameter estimation on control chart properties was given by Jensen et al. (2006). Suppose that we are interested in quickly detecting an increase of size δ opt = 0.5 in the incidence, and that the desired in-control ARL is 500. According to Table 6, the optimal parameters for the RY chart are chosen as: = 0.02 and L = The control limit for the RY chart is h R = L 2 ( π ) = The charting statistic of the RY chart at each time period is obtained according to Equation (8), which is given in Table 9. Figures 2(a) and (b) plot the sample data, X t, and the charting statistic, R Y t, respectively. From Figure 2(a), it can be observed that the observations during years tend to be 13

14 Table 5: Optimal Parameters for the RY Chart When In-Control ARL is 370 δ opt (, L) ( 0.02, 1.809) ( 0.02, 1.832) ( 0.02, 1.846) ( 0.02, 1.828) ( 0.02, 1.834) ( 0.02, 1.828) ( 0.02, 1.820) ARL min (, L) ( 0.02, 1.809) ( 0.02, 1.832) ( 0.02, 1.846) ( 0.02, 1.828) ( 0.02, 1.834) ( 0.02, 1.828) ( 0.02, 1.820) ARL min (, L) ( 0.02, 1.809) ( 0.04, 2.383) ( 0.06, 2.656) ( 0.06, 2.621) ( 0.06, 2.605) ( 0.06, 2.586) ( 0.06, 2.574) ARL min (, L) ( 0.07, 2.895) ( 0.09, 3.044) ( 0.10, 3.044) ( 0.10, 2.984) ( 0.10, 2.965) ( 0.10, 2.938) ( 0.10, 2.922) ARL min (, L) ( 0.10, 3.223) ( 0.13, 3.340) ( 0.13, 3.246) ( 0.14, 3.226) ( 0.14, 3.208) ( 0.15, 3.223) ( 0.15, 3.205) ARL min (, L) ( 0.14, 3.536) ( 0.16, 3.513) ( 0.18, 3.497) ( 0.20, 3.486) ( 0.20, 3.463) ( 0.20, 3.422) ( 0.20, 3.400) ARL min (, L) ( 0.17, 3.716) ( 0.19, 3.660) ( 0.20, 3.574) ( 0.23, 3.594) ( 0.23, 3.568) ( 0.23, 3.526) ( 0.23, 3.501) ARL min (, L) ( 0.19, 3.820) ( 0.21, 3.750) ( 0.22, 3.658) ( 0.25, 3.656) ( 0.25, 3.630) ( 0.25, 3.583) ( 0.25, 3.560) ARL min µ 0 14

15 Table 6: Optimal Parameters for the RY Chart When In-Control ARL is 500 δ opt (, L) ( 0.02, 2.047) ( 0.02, 2.063) ( 0.02, 2.074) ( 0.02, 2.055) ( 0.02, 2.047) ( 0.02, 2.039) ( 0.02, 2.031) ARL min (, L) ( 0.02, 2.047) ( 0.02, 2.063) ( 0.03, 2.363) ( 0.03, 2.336) ( 0.03, 2.328) ( 0.03, 2.324) ( 0.03, 2.303) ARL min (, L) ( 0.03, 2.391) ( 0.06, 2.926) ( 0.06, 2.860) ( 0.06, 2.816) ( 0.06, 2.801) ( 0.07, 2.883) ( 0.07, 2.865) ARL min (, L) ( 0.07, 3.117) ( 0.09, 3.249) ( 0.09, 3.164) ( 0.09, 3.100) ( 0.10, 3.153) ( 0.10, 3.129) ( 0.11, 3.173) ARL min (, L) ( 0.10, 3.606) ( 0.12, 3.480) ( 0.12, 3.381) ( 0.12, 3.302) ( 0.14, 3.393) ( 0.15, 3.404) ( 0.15, 3.386) ARL min (, L) ( 0.12, 3.606) ( 0.16, 3.718) ( 0.17, 3.647) ( 0.18, 3.600) ( 0.19, 3.614) ( 0.19, 3.575) ( 0.20, 3.583) ARL min (, L) ( 0.16, 3.880) ( 0.18, 3.826) ( 0.19, 3.739) ( 0.20, 3.674) ( 0.23, 3.755) ( 0.23, 3.709) ( 0.23, 3.683) ARL min (, L) ( 0.19, 4.044) ( 0.22, 3.999) ( 0.22, 3.856) ( 0.23, 3.782) ( 0.24, 3.785) ( 0.24, 3.740) ( 0.24, 3.714) ARL min µ 0 15

16 Table 7: Optimal Parameters for the RY Chart When In-Control ARL is 1,000 δ opt (, L) ( 0.02, 2.548) ( 0.02, 2.548) ( 0.02, 2.543) ( 0.02, 2.510) ( 0.02, 2.503) ( 0.02, 2.492) ( 0.02, 2.480) ARL min (, L) ( 0.02, 2.547) ( 0.02, 2.548) ( 0.03, 2.814) ( 0.03, 2.773) ( 0.03, 2.764) ( 0.03, 2.744) ( 0.04, 2.902) ARL min (, L) ( 0.05, 3.308) ( 0.05, 3.236) ( 0.06, 3.295) ( 0.06, 3.233) ( 0.06, 3.215) ( 0.06, 3.188) ( 0.06, 3.171) ARL min (, L) ( 0.08, 3.719) ( 0.08, 3.601) ( 0.09, 3.592) ( 0.09, 3.511) ( 0.09, 3.492) ( 0.10, 3.528) ( 0.10, 3.507) ARL min (, L) ( 0.10, 3.923) ( 0.11, 3.857) ( 0.11, 3.743) ( 0.13, 3.773) ( 0.13, 3.750) ( 0.12, 3.656) ( 0.13, 3.687) ARL min (, L) ( 0.13, 4.167) ( 0.14, 4.056) ( 0.14, 3.926) ( 0.16, 3.926) ( 0.16, 3.895) ( 0.17, 3.895) ( 0.17, 3.869) ARL min (, L) ( 0.16, 4.360) ( 0.19, 4.328) ( 0.19, 4.162) ( 0.20, 4.087) ( 0.20, 4.058) ( 0.21, 4.049) ( 0.21, 4.108) ARL min (, L) ( 0.19, 4.532) ( 0.23, 4.498) ( 0.24, 4.357) ( 0.24, 4.230) ( 0.24, 4.196) ( 0.24, 4.142) ( 0.25, 4.141) ARL min µ 0 16

17 Table 8: Optimal Parameters for the RY Chart When In-Control ARL is 1,500 δ opt (, L) ( 0.02, 2.810) ( 0.02, 2.814) ( 0.02, 2.784) ( 0.02, 2.758) ( 0.02, 2.748) ( 0.02, 2.727) ( 0.02, 2.719) ARL min (, L) ( 0.02, 2.810) ( 0.02, 2.814) ( 0.03, 3.056) ( 0.03, 3.010) ( 0.03, 2.996) ( 0.03, 2.978) ( 0.03, 2.961) ARL min (, L) ( 0.05, 3.562) ( 0.05, 3.481) ( 0.05, 3.401) ( 0.05, 3.337) ( 0.06, 3.442) ( 0.06, 3.412) ( 0.06, 3.391) ARL min (, L) ( 0.07, 3.858) ( 0.08, 3.848) ( 0.08, 3.738) ( 0.09, 3.736) ( 0.09, 3.714) ( 0.09, 3.677) ( 0.09, 3.656) ARL min (, L) ( 0.09, 4.088) ( 0.10, 4.025) ( 0.12, 4.040) ( 0.12, 3.940) ( 0.12, 3.915) ( 0.12, 3.871) ( 0.12, 3.846) ARL min (, L) ( 0.13, 4.429) ( 0.14, 4.305) ( 0.14, 4.159) ( 0.16, 4.153) ( 0.16, 4.122) ( 0.16, 4.075) ( 0.16, 4.047) ARL min (, L) ( 0.15, 4.568) ( 0.17, 4.480) ( 0.19, 4.409) ( 0.19, 4.282) ( 0.19, 4.209) ( 0.20, 4.230) ( 0.20, 4.200) ARL min (, L) ( 0.17, 4.692) ( 0.19, 4.579) ( 0.21, 4.489) ( 0.22, 4.393) ( 0.22, 4.358) ( 0.23, 4.335) ( 0.23, 4.300) ARL min µ 0 17

18 Table 9: An Illustrative Example for the RY Chart Year X t Y t Rt Y Year X t Y t Rt Y

19 8 0.4 Cancer Incidence Per 100, Year Year (a) (b) Figure 2: The Annual Incidence Data (a) and the RY Chart (b) larger than the first 16 observations. This exhibits the pattern of an increase in the incidence. From Figure 2(b), it is clear that the RY chart triggers an out-of-control signal in year The RY charting statistic remains above the control limit for the rest of observations and starts the increase since year Therefore, the search for assignable cause can be traced back to year Concluding Remarks This paper extends one-sided EWMA charts from the case of monitoring normal process means to the case of monitoring Poisson data. A Markov chain model is established to analyze the ARL performance for these charts. Both cases with and without normalizing transformation applied to the Poisson data are investigated. The results show that resetting of the current observation or the transformed observation would improve the performance of detecting increases in the Poisson rate, compared to the conventional way of resetting the EWMA statistic to the target. Although the normalization transformation makes the Poisson data much closer to normal, it results in slightly worse performance than the Poisson EWMA chart without transformation. Therefore we suggest using the original observations in the new EWMA Poisson chart. When designing the one-sided Poisson EWMA procedure in our paper, the in-control mean, µ 0, is assumed to be known. In practice; however, it would be more common that the incontrol process parameter of the Poisson distribution is unknown and needs to be estimated. 19

20 The estimation error may have considerable effects on the performance of control charts (Jensen et al. 2006). To make a control chart to perform like the case with known parameter(s), a sufficiently large sample size is often required during a Phase I analysis. In parallel with the work of Jensen et al. (2006), one can determine the appropriate sample size required for establishing the one-sided Poisson EWMA chart. This will be further discussed in another paper. Although the discussions in this paper aim at quick detection of upward changes in Poisson rate, one can easily extend the suggested chart to the case for detecting downward changes. The lower Poisson EWMA chart can be obtained based on the same logic by truncating the observation to the target whenever it is larger than the target. Lucas and Saccucci (1990) have suggested some enhancements for the two-sided EWMA charts, including a combined Shewhart- EWMA and the FIR feature. Although these enhancements are not presented in this paper, it is straightforward to extend them to the one-sided Poisson EWMA charts. Finally, due to the increasing interests in health surveillance, it would be interesting to extend the proposed one-sided EWMA scheme to monitor other attribute data such as Binomial distributions in the future research. Appendix A: Approximation of ARLs for the EZ and EY Charts Using Markov Chain Suppose the in-control region [0, h E ] of the EZ charting statistic is divided into m subintervals, which are labelled as i = 1, 2,, m. The width of each interval is ω = 2h E /(2m 1) except the first is ω/2. Let P (i, j) be the transition probability of Et Z from state i to state j. Define d 1 = 1 [j 1.5 (1 )(i 1)]ω + µ µ 0 d 2 = 1 [j 0.5 (1 )(i 1)]ω + µ µ 0 Then for i = 1, 2,..., m and j 1, P (i, j) = Pr{ E Z t in state j E Z t 1 in state i} = Pr{(j 1)ω 0.5ω < Z t + (1 )(i 1)ω (j 1)ω + 0.5ω [j 1.5 (1 )(i 1)]ω = Pr{ = Pr{d 1 < X t + 3/8 d 2 } < Z t [j 0.5 (1 )(i 1)]ω } 20

21 = 0, d 2 < F (d , µ), F (d , µ) F (d , µ), d 2 d and d 1 < 3 8 and d , where F (, µ) is the cumulative distribution function of a Poisson random variable with mean µ. For j = 1, P (i, j) = Pr{ X t d 2} = F (d , µ). Although the Poisson random variable takes only nonnegative integer values, the left and right sides of the above inequalities in general will not be integers. Denote T the run length of the EZ chart. Then, the probabilities that Et Z goes from one transient state to another state in t steps are given by the matrix R t. Hence, the probability mass function (PMF) of the run length follows Pr(T = t E0 Z ) = p T ini(r t 1 R t ) 1, (9) where E0 Z is an initial value, p ini is the probability vector corresponding to E0 Z, and 1 is a column vector of ones. The zero-state or initial-state ARL is computed by ARL = p T ini (I R) 1 1. (10) For the EZ chart, E0 Z = 0. To compute the steady-state ARL of a control chart using Markov chain approach, one needs to compute the steady-state probability vector. However, the exact stationary probability vector does not exist because the transition probability matrix is not ergodic. To overcome this problem, Lucas and Saccucci (1990) suggested using a cyclical steady-state probability vector as an alternative. It is obtained by adjusting the transition probability matrix so that the control statistic was reset to the initial state whenever it goes into the out-of-control state, namely, P = R (I R) Let p ss be the cyclical steady-state probability vector. Then, p ss is the solution to the system p = P T p subject to 1 T p =1. The cyclical steady-state ARL is given by ARL = p T ss(i R)

22 Analogously, this method was employed to compute the steady-state ARL for the EZ chart in this paper. The zero-state and steady-state ARLs for other control controls can be obtained similarly except that the transition probability matrices need to be redefined. The number of states m may usually be chosen as 30 to evaluate the run length performance, as recommended by Brook and Evans (1972). However, we prefer a much larger m value because this may improve the overall accuracy while affecting the total execution time only marginally, given modern computing power available today. In this paper, we have used the value of m up to m = 300, and the execution time is in the order of a few seconds. Define F (, µ) and ω as above. The transition probability matrix for the EY chart can be obtained as follows: for j 1, P (i, j) = Pr{ E Y t in state j E Y t 1 in state i} = Pr{(j 1)ω 0.5ω < Y t + (1 )(i 1)ω (j 1)ω + 0.5ω [j 1.5 (1 )(i 1)]ω = Pr{ = F (c 2, µ) F (c 1, µ) < Y t [j 0.5 (1 )(i 1)]ω } where c 1 = c 2 = [j 1.5 (1 )(i 1)]ω µ0 + µ 0 [j 0.5 (1 )(i 1)]ω µ0 + µ 0. For j = 1, P (i, j) = Pr{Y t c 2 } = F (c 2, µ). Appendix B: The Markov Chain Model for the RZ and RY Charts Note that the charting statistic of the RZ chart can be written as Since Z + t easy to show that t 1 Rt Z = (1 ) j (Z t j + 1 ) + (1 ) t R0 Z. 2π j=0 0 and t 1 j=0 (1 )j = 1 (1 ) t approaches one as t + for 0 < 1, it is R Z t 1 2π [1 (1 ) t ] 1 2π. 22

23 Thus, the in-control region of Rt Z can be approximated by [ 1/ 2π, h R ], and the interval (h R, + ) is considered to be an absorbing state. The interval [ 1/ 2π, h R ] was partitioned into m sub-intervals, each of width = (h R + 1/ 2π)/m. The mid value for these sub-intervals is 1/ 2π+ (i 0.5) for i = 1, 2,, m. For i = 1, 2,..., m and j = 1, 2,..., m, the transition probability for the charting statistic Rt Z from state i to state j is P (i, j) = Pr{ R Z t in state j R Z t 1 in state i} = Pr{ 1 2π + (j 1) < (Z + t 1 2π ) + (1 )( 1 2π + (i 0.5) ) 1 2π + j } [j 1 (1 )(i 0.5)] = Pr{ < Z t + [j (1 )(i 0.5)] }. Define and then a 1 = a 2 = [j 1 (1 )(i 0.5)], [j (1 )(i 0.5)], P (i, j) = Pr{a 1 < Z t + a 2 }, 0, a 2 < 0 = F (b 2, µ), a 2 0 and a 1 < 0 F (b 2, µ) F (b 1, µ), a 2 0 and a 1 0, where b 1 = b 2 = ( ) 2 0.5a 1 + µ µ 0 8, ( ) 2 0.5a 2 + µ µ 0 8. The transition probability for the RY chart can be calculated without modifications except that values of b 1 and b 2 are redefined as b 1 = a 1 µ0 + µ 0, b 2 = a 2 µ0 + µ 0. 23

24 References [1] Anscombe, F.J. (1948) The transformation of poisson, binomial, and negative binomial data. Biometrika, 35, [2] Freeman, M.F. and Tukey, J.W. (1950) Transformations related to the angular and the square root. Annals of Mathematical Statistics, 21, [3] Barr, D.R. and Sherrill, E.T. (1999) Mean and variance of truncated normal distributions. The American Statistician, 53, [4] Borror, C.M., Champ, C.W. and Rigdon, S.E. (1998) Poisson EWMA control charts. Journal of Quality Technology, 30, [5] Brook, D. and Evans, D.A. (1972) An approach to the probability distribution of CUSUM run length. Biometrika, 59, [6] Champ, C.W., Woodall, W.H. and Mohsen, H.A. (1991) A generalized quality control procedure. Statistics and Probability Letters, 11, [7] Chan, L.Y., Ouyang, J.T. and Lau, Y.K. (2007) A two-stage cumulative quality control chart for monitoring poisson processes. Journal of Quality Technology, 39, [8] Crowder, S.V. and Hamilton, M. (1992) Average run lengths of EWMA controls for monitoring a process standard deviation. Journal of Quality Technology, 24, [9] Gan, F.F. (1990) Monitoring poisson observations using modified exponentially weighted moving average cnotrol charts. Communications in Statistics: Simulation and Computation, 19, [10] Gan, F.F. (1998) Designs of one- and two-sided exponential EWMA charts. Journal of Quality Technology, 30, [11] Jensen, W.A., Jones-Farmer, L.A., Champ, C.W., Woodall, W.H. (2006) Effects of parameter estimation on control chart properties: a literature review. Journal of Quality Technology, 38: [12] Lucas, J.M. and Crosier, R.B. (1982) Fast initial response for CUSUM quality-control schemes: give your CUSUM a head start. Technometrics, 24,

25 [13] Lucas, J.M. (1985) Counted data CUSUm s. Technometrics, 27, [14] Lucas, J.M. and Saccucci, M.S. (1990) Exponentially weighted moving average control schemes: properties and enhancements. Technometrics, 32, [15] Martz, H.F. and Kvam, P.H. (1996) Detecting trends and patterns in reliability data over time using exponentially weighted moving-averages. Reliability Engineering and System Safety, 51, [16] Mei, Y.J., Han, S.W., Tsui, K.L. (2010) Early detection of a change in Poisson rate after accounting for population size effects. Statistica Sinica, to be published. [17] Montgomery, D.C. (2005). Introduction to Statistical Quality Control, 5th ed. New York: Wiley. [18] Quesenberry, C.P. (1991a) SPC Q charts for start-up processes and short and long runs. Journal of Quality Technology, 23, [19] Quesenberry, C.P. (1991b) SPC Q charts for a binomial parameter: short and long Runs. Journal of Quality Technology, 23, [20] Quesenberry, C.P. (1991c). SPC Q charts for a poisson parameter : short and long runs. Journal of Quality Technology, 23, [21] Rossi, G., Lampugnani, L. and March, M. (1999) An approximate CUSUM procedure for surveillance of health events. Statistics in Medicine, 18, [22] Ryan, T.P. and Schwertman, N.C. (1997) Optimal limits for attributes control charts. Journal of Quality Technology, 29, [23] Shu, L., Jiang, W. and Wu, S. (2007). A One-Sided EWMA Control Chart for Monitoring Process Means. Communication in Statistics: Simulation and Computation 36, [24] Shu, L. and Jiang, W. (2008) A new EWMA chart for monitoring process dispersion. Journal of Quality Technology, 40, [25] Tsui, K.L., Chiu, W., Gierlich, P., Goldsman, D., Liu, X. and Maschek, T. (2008) A Review of health-care, public health, and syndromic surveyllance. Quality Engineering, 20,

26 [26] White, C.H. and Keats, J.B. (1996) ARLs and high-order run-length moments for the poisson CUSUM. Journal of Quality Technology, 28, [27] White, C.H., Keats, J.B. and Stanley, J. (1997) Poisson CUSUM vs. c-chart for defect rate. Quality Engineering, 9, [28] Woodall, W.H. (1997) Control charts based on attribute data: bibliography and review. Journal of Quality Technology, 29, [29] Woodall, W.H. (2006) The use of control charts in health-care and public health surveillance. Journal of Quality Technology, 38,

An Adaptive Exponentially Weighted Moving Average Control Chart for Monitoring Process Variances

An Adaptive Exponentially Weighted Moving Average Control Chart for Monitoring Process Variances An Adaptive Exponentially Weighted Moving Average Control Chart for Monitoring Process Variances Lianjie Shu Faculty of Business Administration University of Macau Taipa, Macau (ljshu@umac.mo) Abstract

More information

CUMULATIVE SUM CHARTS FOR HIGH YIELD PROCESSES

CUMULATIVE SUM CHARTS FOR HIGH YIELD PROCESSES Statistica Sinica 11(2001), 791-805 CUMULATIVE SUM CHARTS FOR HIGH YIELD PROCESSES T. C. Chang and F. F. Gan Infineon Technologies Melaka and National University of Singapore Abstract: The cumulative sum

More information

A Modified Poisson Exponentially Weighted Moving Average Chart Based on Improved Square Root Transformation

A Modified Poisson Exponentially Weighted Moving Average Chart Based on Improved Square Root Transformation Thailand Statistician July 216; 14(2): 197-22 http://statassoc.or.th Contributed paper A Modified Poisson Exponentially Weighted Moving Average Chart Based on Improved Square Root Transformation Saowanit

More information

The occurrence of rare events in manufacturing processes, e.g. nonconforming items or machine failures, is frequently modeled

The occurrence of rare events in manufacturing processes, e.g. nonconforming items or machine failures, is frequently modeled Research Article (wileyonlinelibrary.com) DOI: 1.12/qre.1495 Published online in Wiley Online Library Exponential CUSUM Charts with Estimated Control Limits Min Zhang, a Fadel M. Megahed b * and William

More information

A Theoretically Appropriate Poisson Process Monitor

A Theoretically Appropriate Poisson Process Monitor International Journal of Performability Engineering, Vol. 8, No. 4, July, 2012, pp. 457-461. RAMS Consultants Printed in India A Theoretically Appropriate Poisson Process Monitor RYAN BLACK and JUSTIN

More information

Likelihood-Based EWMA Charts for Monitoring Poisson Count Data with Time-Varying Sample Sizes

Likelihood-Based EWMA Charts for Monitoring Poisson Count Data with Time-Varying Sample Sizes Likelihood-Based EWMA Charts for Monitoring Poisson Count Data with Time-Varying Sample Sizes Qin Zhou 1,3, Changliang Zou 1, Zhaojun Wang 1 and Wei Jiang 2 1 LPMC and Department of Statistics, School

More information

EFFICIENT CHANGE DETECTION METHODS FOR BIO AND HEALTHCARE SURVEILLANCE

EFFICIENT CHANGE DETECTION METHODS FOR BIO AND HEALTHCARE SURVEILLANCE EFFICIENT CHANGE DETECTION METHODS FOR BIO AND HEALTHCARE SURVEILLANCE A Thesis Presented to The Academic Faculty by Sung Won Han In Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

More information

THE DETECTION OF SHIFTS IN AUTOCORRELATED PROCESSES WITH MR AND EWMA CHARTS

THE DETECTION OF SHIFTS IN AUTOCORRELATED PROCESSES WITH MR AND EWMA CHARTS THE DETECTION OF SHIFTS IN AUTOCORRELATED PROCESSES WITH MR AND EWMA CHARTS Karin Kandananond, kandananond@hotmail.com Faculty of Industrial Technology, Rajabhat University Valaya-Alongkorn, Prathumthani,

More information

A problem faced in the context of control charts generally is the measurement error variability. This problem is the result of the inability to

A problem faced in the context of control charts generally is the measurement error variability. This problem is the result of the inability to A problem faced in the context of control charts generally is the measurement error variability. This problem is the result of the inability to measure accurately the variable of interest X. The use of

More information

Analysis and Design of One- and Two-Sided Cusum Charts with Known and Estimated Parameters

Analysis and Design of One- and Two-Sided Cusum Charts with Known and Estimated Parameters Georgia Southern University Digital Commons@Georgia Southern Electronic Theses & Dissertations Graduate Studies, Jack N. Averitt College of Spring 2007 Analysis and Design of One- and Two-Sided Cusum Charts

More information

The Robustness of the Multivariate EWMA Control Chart

The Robustness of the Multivariate EWMA Control Chart The Robustness of the Multivariate EWMA Control Chart Zachary G. Stoumbos, Rutgers University, and Joe H. Sullivan, Mississippi State University Joe H. Sullivan, MSU, MS 39762 Key Words: Elliptically symmetric,

More information

Optimal SPRT and CUSUM Procedures using Compressed Limit Gauges

Optimal SPRT and CUSUM Procedures using Compressed Limit Gauges Optimal SPRT and CUSUM Procedures using Compressed Limit Gauges P. Lee Geyer Stefan H. Steiner 1 Faculty of Business McMaster University Hamilton, Ontario L8S 4M4 Canada Dept. of Statistics and Actuarial

More information

CONTROL CHARTS FOR THE GENERALIZED POISSON PROCESS WITH UNDER-DISPERSION

CONTROL CHARTS FOR THE GENERALIZED POISSON PROCESS WITH UNDER-DISPERSION International J. of Math. Sci. & Engg. Appls. (IJMSEA) ISSN 0973-9424, Vol. 10 No. III (December, 2016), pp. 173-181 CONTROL CHARTS FOR THE GENERALIZED POISSON PROCESS WITH UNDER-DISPERSION NARUNCHARA

More information

A New Demerit Control Chart for Monitoring the Quality of Multivariate Poisson Processes. By Jeh-Nan Pan Chung-I Li Min-Hung Huang

A New Demerit Control Chart for Monitoring the Quality of Multivariate Poisson Processes. By Jeh-Nan Pan Chung-I Li Min-Hung Huang Athens Journal of Technology and Engineering X Y A New Demerit Control Chart for Monitoring the Quality of Multivariate Poisson Processes By Jeh-Nan Pan Chung-I Li Min-Hung Huang This study aims to develop

More information

An Investigation of Combinations of Multivariate Shewhart and MEWMA Control Charts for Monitoring the Mean Vector and Covariance Matrix

An Investigation of Combinations of Multivariate Shewhart and MEWMA Control Charts for Monitoring the Mean Vector and Covariance Matrix Technical Report Number 08-1 Department of Statistics Virginia Polytechnic Institute and State University, Blacksburg, Virginia January, 008 An Investigation of Combinations of Multivariate Shewhart and

More information

Synthetic and runs-rules charts combined with an chart: Theoretical discussion

Synthetic and runs-rules charts combined with an chart: Theoretical discussion Synthetic and runs-rules charts combined with an chart: Theoretical discussion Sandile C. Shongwe and Marien A. Graham Department of Statistics University of Pretoria South Africa sandile.shongwe@up.ac.za

More information

Quality Control & Statistical Process Control (SPC)

Quality Control & Statistical Process Control (SPC) Quality Control & Statistical Process Control (SPC) DR. RON FRICKER PROFESSOR & HEAD, DEPARTMENT OF STATISTICS DATAWORKS CONFERENCE, MARCH 22, 2018 Agenda Some Terminology & Background SPC Methods & Philosophy

More information

Surveillance of Infectious Disease Data using Cumulative Sum Methods

Surveillance of Infectious Disease Data using Cumulative Sum Methods Surveillance of Infectious Disease Data using Cumulative Sum Methods 1 Michael Höhle 2 Leonhard Held 1 1 Institute of Social and Preventive Medicine University of Zurich 2 Department of Statistics University

More information

Self-Starting Control Chart for Simultaneously Monitoring Process Mean and Variance

Self-Starting Control Chart for Simultaneously Monitoring Process Mean and Variance International Journal of Production Research Vol. 00, No. 00, 15 March 2008, 1 14 Self-Starting Control Chart for Simultaneously Monitoring Process Mean and Variance Zhonghua Li a, Jiujun Zhang a,b and

More information

Weighted Likelihood Ratio Chart for Statistical Monitoring of Queueing Systems

Weighted Likelihood Ratio Chart for Statistical Monitoring of Queueing Systems Weighted Likelihood Ratio Chart for Statistical Monitoring of Queueing Systems Dequan Qi 1, Zhonghua Li 2, Xuemin Zi 3, Zhaojun Wang 2 1 LPMC and School of Mathematical Sciences, Nankai University, Tianjin

More information

Performance Analysis of Queue Length Monitoring of M/G/1 Systems

Performance Analysis of Queue Length Monitoring of M/G/1 Systems Performance Analysis of Queue Length Monitoring of M/G/1 Systems Nan Chen, 1 Yuan Yuan, 2 Shiyu Zhou 2 1 Department of Industrial and Systems Engineering, National University of Singapore, Singapore 2

More information

Directionally Sensitive Multivariate Statistical Process Control Methods

Directionally Sensitive Multivariate Statistical Process Control Methods Directionally Sensitive Multivariate Statistical Process Control Methods Ronald D. Fricker, Jr. Naval Postgraduate School October 5, 2005 Abstract In this paper we develop two directionally sensitive statistical

More information

Module B1: Multivariate Process Control

Module B1: Multivariate Process Control Module B1: Multivariate Process Control Prof. Fugee Tsung Hong Kong University of Science and Technology Quality Lab: http://qlab.ielm.ust.hk I. Multivariate Shewhart chart WHY MULTIVARIATE PROCESS CONTROL

More information

An Economic Alternative to the c Chart

An Economic Alternative to the c Chart University of Arkansas, Fayetteville ScholarWorks@UARK Theses and Dissertations 12-2012 An Economic Alternative to the c Chart Ryan William Black University of Arkansas, Fayetteville Follow this and additional

More information

CONTROL CHARTS FOR MULTIVARIATE NONLINEAR TIME SERIES

CONTROL CHARTS FOR MULTIVARIATE NONLINEAR TIME SERIES REVSTAT Statistical Journal Volume 13, Number, June 015, 131 144 CONTROL CHARTS FOR MULTIVARIATE NONLINEAR TIME SERIES Authors: Robert Garthoff Department of Statistics, European University, Große Scharrnstr.

More information

Distribution-Free Monitoring of Univariate Processes. Peihua Qiu 1 and Zhonghua Li 1,2. Abstract

Distribution-Free Monitoring of Univariate Processes. Peihua Qiu 1 and Zhonghua Li 1,2. Abstract Distribution-Free Monitoring of Univariate Processes Peihua Qiu 1 and Zhonghua Li 1,2 1 School of Statistics, University of Minnesota, USA 2 LPMC and Department of Statistics, Nankai University, China

More information

Faculty of Science and Technology MASTER S THESIS

Faculty of Science and Technology MASTER S THESIS Faculty of Science and Technology MASTER S THESIS Study program/ Specialization: Spring semester, 20... Open / Restricted access Writer: Faculty supervisor: (Writer s signature) External supervisor(s):

More information

Mathematical and Computer Modelling. Economic design of EWMA control charts based on loss function

Mathematical and Computer Modelling. Economic design of EWMA control charts based on loss function Mathematical and Computer Modelling 49 (2009) 745 759 Contents lists available at ScienceDirect Mathematical and Computer Modelling journal homepage: www.elsevier.com/locate/mcm Economic design of EWMA

More information

Variations in a manufacturing process can be categorized into common cause and special cause variations. In the presence of

Variations in a manufacturing process can be categorized into common cause and special cause variations. In the presence of Research Article (wileyonlinelibrary.com) DOI: 10.1002/qre.1514 Published online in Wiley Online Library Memory-Type Control Charts for Monitoring the Process Dispersion Nasir Abbas, a * Muhammad Riaz

More information

Zero-Inflated Models in Statistical Process Control

Zero-Inflated Models in Statistical Process Control Chapter 6 Zero-Inflated Models in Statistical Process Control 6.0 Introduction In statistical process control Poisson distribution and binomial distribution play important role. There are situations wherein

More information

A new multivariate CUSUM chart using principal components with a revision of Crosier's chart

A new multivariate CUSUM chart using principal components with a revision of Crosier's chart Title A new multivariate CUSUM chart using principal components with a revision of Crosier's chart Author(s) Chen, J; YANG, H; Yao, JJ Citation Communications in Statistics: Simulation and Computation,

More information

The Efficiency of the VSI Exponentially Weighted Moving Average Median Control Chart

The Efficiency of the VSI Exponentially Weighted Moving Average Median Control Chart The Efficiency of the VSI Exponentially Weighted Moving Average Median Control Chart Kim Phuc Tran, Philippe Castagliola, Thi-Hien Nguyen, Anne Cuzol To cite this version: Kim Phuc Tran, Philippe Castagliola,

More information

DIAGNOSIS OF BIVARIATE PROCESS VARIATION USING AN INTEGRATED MSPC-ANN SCHEME

DIAGNOSIS OF BIVARIATE PROCESS VARIATION USING AN INTEGRATED MSPC-ANN SCHEME DIAGNOSIS OF BIVARIATE PROCESS VARIATION USING AN INTEGRATED MSPC-ANN SCHEME Ibrahim Masood, Rasheed Majeed Ali, Nurul Adlihisam Mohd Solihin and Adel Muhsin Elewe Faculty of Mechanical and Manufacturing

More information

Songklanakarin Journal of Science and Technology SJST R1 Sukparungsee

Songklanakarin Journal of Science and Technology SJST R1 Sukparungsee Songklanakarin Journal of Science and Technology SJST-0-0.R Sukparungsee Bivariate copulas on the exponentially weighted moving average control chart Journal: Songklanakarin Journal of Science and Technology

More information

Modified cumulative sum quality control scheme

Modified cumulative sum quality control scheme Journal of Engineering and Technology Research Vol. 2(12), pp. 226-236, December 2010 Available online at http:// www.academicjournals.org/jetr ISSN 2006-9790 2010 Academic Journals Full Length Research

More information

SCIENCE & TECHNOLOGY

SCIENCE & TECHNOLOGY Pertanika J. Sci. & Technol. 24 (1): 177-189 (2016) SCIENCE & TECHNOLOGY Journal homepage: http://www.pertanika.upm.edu.my/ A Comparative Study of the Group Runs and Side Sensitive Group Runs Control Charts

More information

Detecting Assignable Signals via Decomposition of MEWMA Statistic

Detecting Assignable Signals via Decomposition of MEWMA Statistic International Journal of Mathematics and Statistics Invention (IJMSI) E-ISSN: 2321 4767 P-ISSN: 2321-4759 Volume 4 Issue 1 January. 2016 PP-25-29 Detecting Assignable Signals via Decomposition of MEWMA

More information

Monitoring aggregated Poisson data with probability control limits

Monitoring aggregated Poisson data with probability control limits XXV Simposio Internacional de Estadística 2015 Armenia, Colombia, 5, 6, 7 y 8 de Agosto de 2015 Monitoring aggregated Poisson data with probability control limits Victor Hugo Morales Ospina 1,a, José Alberto

More information

AN ANALYSIS OF SHEWHART QUALITY CONTROL CHARTS TO MONITOR BOTH THE MEAN AND VARIABILITY KEITH JACOB BARRS. A Research Project Submitted to the Faculty

AN ANALYSIS OF SHEWHART QUALITY CONTROL CHARTS TO MONITOR BOTH THE MEAN AND VARIABILITY KEITH JACOB BARRS. A Research Project Submitted to the Faculty AN ANALYSIS OF SHEWHART QUALITY CONTROL CHARTS TO MONITOR BOTH THE MEAN AND VARIABILITY by KEITH JACOB BARRS A Research Project Submitted to the Faculty of the College of Graduate Studies at Georgia Southern

More information

Control Charts for Monitoring the Zero-Inflated Generalized Poisson Processes

Control Charts for Monitoring the Zero-Inflated Generalized Poisson Processes Thai Journal of Mathematics Volume 11 (2013) Number 1 : 237 249 http://thaijmath.in.cmu.ac.th ISSN 1686-0209 Control Charts for Monitoring the Zero-Inflated Generalized Poisson Processes Narunchara Katemee

More information

THE CUSUM MEDIAN CHART FOR KNOWN AND ESTIMATED PARAMETERS

THE CUSUM MEDIAN CHART FOR KNOWN AND ESTIMATED PARAMETERS THE CUSUM MEDIAN CHART FOR KNOWN AND ESTIMATED PARAMETERS Authors: Philippe Castagliola Université de Nantes & LS2N UMR CNRS 6004, Nantes, France (philippe.castagliola@univ-nantes.fr) Fernanda Otilia Figueiredo

More information

On the performance of Shewhart-type synthetic and runs-rules charts combined with an chart

On the performance of Shewhart-type synthetic and runs-rules charts combined with an chart On the performance of Shewhart-type synthetic and runs-rules charts combined with an chart Sandile Charles Shongwe and Marien Alet Graham Department of Statistics University of Pretoria South Africa Abstract

More information

Robustness of the EWMA control chart for individual observations

Robustness of the EWMA control chart for individual observations 1 Robustness of the EWMA control chart for individual observations S.W. Human Department of Statistics University of Pretoria Lynnwood Road, Pretoria, South Africa schalk.human@up.ac.za P. Kritzinger Department

More information

Multivariate Charts for Multivariate. Poisson-Distributed Data. Busaba Laungrungrong

Multivariate Charts for Multivariate. Poisson-Distributed Data. Busaba Laungrungrong Multivariate Charts for Multivariate Poisson-Distributed Data by Busaba Laungrungrong A Dissertation Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy Approved November

More information

Early Detection of a Change in Poisson Rate After Accounting For Population Size Effects

Early Detection of a Change in Poisson Rate After Accounting For Population Size Effects Early Detection of a Change in Poisson Rate After Accounting For Population Size Effects School of Industrial and Systems Engineering, Georgia Institute of Technology, 765 Ferst Drive NW, Atlanta, GA 30332-0205,

More information

Detection and Diagnosis of Unknown Abrupt Changes Using CUSUM Multi-Chart Schemes

Detection and Diagnosis of Unknown Abrupt Changes Using CUSUM Multi-Chart Schemes Sequential Analysis, 26: 225 249, 2007 Copyright Taylor & Francis Group, LLC ISSN: 0747-4946 print/532-476 online DOI: 0.080/0747494070404765 Detection Diagnosis of Unknown Abrupt Changes Using CUSUM Multi-Chart

More information

Control charts are used for monitoring the performance of a quality characteristic. They assist process

Control charts are used for monitoring the performance of a quality characteristic. They assist process QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL Qual. Reliab. Engng. Int. 2009; 25:875 883 Published online 3 March 2009 in Wiley InterScience (www.interscience.wiley.com)..1007 Research Identifying

More information

Jumps in binomial AR(1) processes

Jumps in binomial AR(1) processes Jumps in binomial AR1 processes Christian H. Weiß To cite this version: Christian H. Weiß. Jumps in binomial AR1 processes. Statistics and Probability Letters, Elsevier, 009, 79 19, pp.01. .

More information

Nonparametric Multivariate Control Charts Based on. A Linkage Ranking Algorithm

Nonparametric Multivariate Control Charts Based on. A Linkage Ranking Algorithm Nonparametric Multivariate Control Charts Based on A Linkage Ranking Algorithm Helen Meyers Bush Data Mining & Advanced Analytics, UPS 55 Glenlake Parkway, NE Atlanta, GA 30328, USA Panitarn Chongfuangprinya

More information

The Efficiency of the 4-out-of-5 Runs Rules Scheme for monitoring the Ratio of Population Means of a Bivariate Normal distribution

The Efficiency of the 4-out-of-5 Runs Rules Scheme for monitoring the Ratio of Population Means of a Bivariate Normal distribution Proceedings of the 22nd ISSAT International Conference on Reliability and Quality in Design August 4-6, 2016 - Los Angeles, California, U.S.A. The Efficiency of the 4-out-of-5 Runs Rules Scheme for monitoring

More information

JINHO KIM ALL RIGHTS RESERVED

JINHO KIM ALL RIGHTS RESERVED 014 JINHO KIM ALL RIGHTS RESERVED CHANGE POINT DETECTION IN UNIVARIATE AND MULTIVARIATE PROCESSES by JINHO KIM A dissertation submitted to the Graduate School-New Brunswick Rutgers, The State University

More information

Because of the global market that considers customer satisfaction as a primary objective, quality has been established as a key

Because of the global market that considers customer satisfaction as a primary objective, quality has been established as a key Synthetic Phase II Shewhart-type Attributes Control Charts When Process Parameters are Estimated Philippe Castagliola, a * Shu Wu, a,b Michael B. C. Khoo c and S. Chakraborti d,e The performance of attributes

More information

On ARL-unbiased c-charts for i.i.d. and INAR(1) Poisson counts

On ARL-unbiased c-charts for i.i.d. and INAR(1) Poisson counts On ARL-unbiased c-charts for iid and INAR(1) Poisson counts Manuel Cabral Morais (1) with Sofia Paulino (2) and Sven Knoth (3) (1) Department of Mathematics & CEMAT IST, ULisboa, Portugal (2) IST, ULisboa,

More information

A multivariate exponentially weighted moving average control chart for monitoring process variability

A multivariate exponentially weighted moving average control chart for monitoring process variability Journal of Applied Statistics, Vol. 30, No. 5, 2003, 507 536 A multivariate exponentially weighted moving average control chart for monitoring process variability ARTHUR B. YEH 1, DENNIS K. J. LIN 2, HONGHONG

More information

Monitoring Expense Report Errors: Control Charts Under Independence and Dependence. Darren Williams. (Under the direction of Dr.

Monitoring Expense Report Errors: Control Charts Under Independence and Dependence. Darren Williams. (Under the direction of Dr. Monitoring Expense Report Errors: Control Charts Under Independence and Dependence by Darren Williams (Under the direction of Dr. Lynne Seymour) Abstract Control charts were devised to evaluate offices

More information

2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008

2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008 MIT OpenCourseWare http://ocw.mit.edu 2.830J / 6.780J / ESD.63J Control of Processes (SMA 6303) Spring 2008 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/term

More information

Cumulative probability control charts for geometric and exponential process characteristics

Cumulative probability control charts for geometric and exponential process characteristics int j prod res, 2002, vol 40, no 1, 133±150 Cumulative probability control charts for geometric and exponential process characteristics L Y CHANy*, DENNIS K J LINz, M XIE} and T N GOH} A statistical process

More information

Performance of Conventional X-bar Chart for Autocorrelated Data Using Smaller Sample Sizes

Performance of Conventional X-bar Chart for Autocorrelated Data Using Smaller Sample Sizes , 23-25 October, 2013, San Francisco, USA Performance of Conventional X-bar Chart for Autocorrelated Data Using Smaller Sample Sizes D. R. Prajapati Abstract Control charts are used to determine whether

More information

A Multivariate EWMA Control Chart for Skewed Populations using Weighted Variance Method

A Multivariate EWMA Control Chart for Skewed Populations using Weighted Variance Method OPEN ACCESS Int. Res. J. of Science & Engineering, 04; Vol. (6): 9-0 ISSN: 3-005 RESEARCH ARTICLE A Multivariate EMA Control Chart for Skewed Populations using eighted Variance Method Atta AMA *, Shoraim

More information

EARLY DETECTION OF A CHANGE IN POISSON RATE AFTER ACCOUNTING FOR POPULATION SIZE EFFECTS

EARLY DETECTION OF A CHANGE IN POISSON RATE AFTER ACCOUNTING FOR POPULATION SIZE EFFECTS Statistica Sinica 21 (2011), 597-624 EARLY DETECTION OF A CHANGE IN POISSON RATE AFTER ACCOUNTING FOR POPULATION SIZE EFFECTS Yajun Mei, Sung Won Han and Kwok-Leung Tsui Georgia Institute of Technology

More information

Statistical quality control (SQC)

Statistical quality control (SQC) Statistical quality control (SQC) The application of statistical techniques to measure and evaluate the quality of a product, service, or process. Two basic categories: I. Statistical process control (SPC):

More information

CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules

CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules. Introduction: The is widely used in industry to monitor the number of fraction nonconforming units. A nonconforming unit is

More information

A STUDY ON IMPROVING THE PERFORMANCE OF CONTROL CHARTS UNDER NON-NORMAL DISTRIBUTIONS SUN TINGTING

A STUDY ON IMPROVING THE PERFORMANCE OF CONTROL CHARTS UNDER NON-NORMAL DISTRIBUTIONS SUN TINGTING A STUDY ON IMPROVING THE PERFORMANCE OF CONTROL CHARTS UNDER NON-NORMAL DISTRIBUTIONS SUN TINGTING NATIONAL UNIVERSITY OF SINGAPORE 004 A STUDY ON IMPROVING THE PERFORMANCE OF CONTROL CHARTS UNDER NON-NORMAL

More information

Statistical process control of the stochastic complexity of discrete processes

Statistical process control of the stochastic complexity of discrete processes UDC 59.84 59.876. S p e c i a l G u e s t I s s u e CDQM, Volume 8, umber, 5, pp. 55-6 COMMUICATIOS I DEPEDABILITY AD QUALITY MAAGEMET An International Journal Statistical process control of the stochastic

More information

Design and Implementation of CUSUM Exceedance Control Charts for Unknown Location

Design and Implementation of CUSUM Exceedance Control Charts for Unknown Location Design and Implementation of CUSUM Exceedance Control Charts for Unknown Location MARIEN A. GRAHAM Department of Statistics University of Pretoria South Africa marien.graham@up.ac.za S. CHAKRABORTI Department

More information

Statistical Process Control for Multivariate Categorical Processes

Statistical Process Control for Multivariate Categorical Processes Statistical Process Control for Multivariate Categorical Processes Fugee Tsung The Hong Kong University of Science and Technology Fugee Tsung 1/27 Introduction Typical Control Charts Univariate continuous

More information

MONITORING BIVARIATE PROCESSES WITH A SYNTHETIC CONTROL CHART BASED ON SAMPLE RANGES

MONITORING BIVARIATE PROCESSES WITH A SYNTHETIC CONTROL CHART BASED ON SAMPLE RANGES Blumenau-SC, 27 a 3 de Agosto de 217. MONITORING BIVARIATE PROCESSES WITH A SYNTHETIC CONTROL CHART BASED ON SAMPLE RANGES Marcela A. G. Machado São Paulo State University (UNESP) Departamento de Produção,

More information

Control charts continue to play a transformative role in all walks of life in the 21st century. The mean and the variance of a

Control charts continue to play a transformative role in all walks of life in the 21st century. The mean and the variance of a Research Article (wileyonlinelibrary.com) DOI: 10.1002/qre.1249 Published online in Wiley Online Library A Distribution-free Control Chart for the Joint Monitoring of Location and Scale A. Mukherjee, a

More information

Control Charts for Zero-Inflated Poisson Models

Control Charts for Zero-Inflated Poisson Models Applied Mathematical Sciences, Vol. 6, 2012, no. 56, 2791-2803 Control Charts for Zero-Inflated Poisson Models Narunchara Katemee Department of Applied Statistics, Faculty of Applied Science King Mongkut

More information

Optimal Design of Second-Order Linear Filters for Control Charting

Optimal Design of Second-Order Linear Filters for Control Charting Optimal Design of Second-Order Linear Filters for Control Charting Chang-Ho CHIN School of Mechanical and Industrial Systems Engineering Kyung Hee University Yongin-si, Gyeonggi-do 446-701 Republic of

More information

A Control Chart for Time Truncated Life Tests Using Exponentiated Half Logistic Distribution

A Control Chart for Time Truncated Life Tests Using Exponentiated Half Logistic Distribution Appl. Math. Inf. Sci. 12, No. 1, 125-131 (2018 125 Applied Mathematics & Information Sciences An International Journal http://dx.doi.org/10.18576/amis/120111 A Control Chart for Time Truncated Life Tests

More information

A NONLINEAR FILTER CONTROL CHART FOR DETECTING DYNAMIC CHANGES

A NONLINEAR FILTER CONTROL CHART FOR DETECTING DYNAMIC CHANGES Statistica Sinica 20 (2010), 1077-1096 A NONLINEAR FILTER CONTROL CHART FOR DETECTING DYNAMIC CHANGES Dong Han 1, Fugee Tsung 2, Yanting Li 1 and Kaibo Wang 3 1 Shanghai Jiao Tong University, 2 Hong Kong

More information

Section II: Assessing Chart Performance. (Jim Benneyan)

Section II: Assessing Chart Performance. (Jim Benneyan) Section II: Assessing Chart Performance (Jim Benneyan) 1 Learning Objectives Understand concepts of chart performance Two types of errors o Type 1: Call an in-control process out-of-control o Type 2: Call

More information

On Efficient Memory-Type Control Charts for Monitoring out of Control Signals in a Process Using Diabetic Data

On Efficient Memory-Type Control Charts for Monitoring out of Control Signals in a Process Using Diabetic Data Biomedical Statistics and Informatics 017; (4): 138-144 http://www.sciencepublishinggroup.com/j/bsi doi: 10.11648/j.bsi.017004.1 On Efficient Memory-Type Control Charts for Monitoring out of Control Signals

More information

Optimal Cusum Control Chart for Censored Reliability Data with Log-logistic Distribution

Optimal Cusum Control Chart for Censored Reliability Data with Log-logistic Distribution CMST 21(4) 221-227 (2015) DOI:10.12921/cmst.2015.21.04.006 Optimal Cusum Control Chart for Censored Reliability Data with Log-logistic Distribution B. Sadeghpour Gildeh, M. Taghizadeh Ashkavaey Department

More information

A Simulation Study Of The Impact Of Forecast Recovery For Control Charts Applied To ARMA Processes

A Simulation Study Of The Impact Of Forecast Recovery For Control Charts Applied To ARMA Processes Journal of Modern Applied Statistical Methods Volume 1 Issue 2 Article 43 11-1-2002 A Simulation Study Of The Impact Of Forecast Recovery For Control Charts Applied To ARMA Processes John N. Dyer Georgia

More information

2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008

2.830J / 6.780J / ESD.63J Control of Manufacturing Processes (SMA 6303) Spring 2008 MIT OpenCourseWare http://ocw.mit.edu 2.830J / 6.780J / ESD.63J Control of Processes (SMA 6303) Spring 2008 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

More information

arxiv: v1 [stat.me] 14 Jan 2019

arxiv: v1 [stat.me] 14 Jan 2019 arxiv:1901.04443v1 [stat.me] 14 Jan 2019 An Approach to Statistical Process Control that is New, Nonparametric, Simple, and Powerful W.J. Conover, Texas Tech University, Lubbock, Texas V. G. Tercero-Gómez,Tecnológico

More information

assumption identically change method. 1. Introduction1 .iust.ac.ir/ ABSTRACT identifying KEYWORDS estimation, Correlation,

assumption identically change method. 1. Introduction1 .iust.ac.ir/   ABSTRACT identifying KEYWORDS estimation, Correlation, International Journal Industrial Engineering & Production Research (207) December 207, Volume 28, Number 4 pp. 367-376 DOI: 0.22068/ijiepr.28.4.367 http://ijiepr..iust.ac.ir/ Change-Point Estimation High-Yiel

More information

Univariate and Multivariate Surveillance Methods for Detecting Increases in Incidence Rates

Univariate and Multivariate Surveillance Methods for Detecting Increases in Incidence Rates Univariate and Multivariate Surveillance Methods for Detecting Increases in Incidence Rates Michael D. Joner, Jr. Dissertation submitted to the Faculty of the Virginia Polytechnic Institute and State University

More information

First Semester Dr. Abed Schokry SQC Chapter 9: Cumulative Sum and Exponential Weighted Moving Average Control Charts

First Semester Dr. Abed Schokry SQC Chapter 9: Cumulative Sum and Exponential Weighted Moving Average Control Charts Department of Industrial Engineering First Semester 2014-2015 Dr. Abed Schokry SQC Chapter 9: Cumulative Sum and Exponential Weighted Moving Average Control Charts Learning Outcomes After completing this

More information

Construction of An Efficient Multivariate Dynamic Screening System. Jun Li a and Peihua Qiu b. Abstract

Construction of An Efficient Multivariate Dynamic Screening System. Jun Li a and Peihua Qiu b. Abstract Construction of An Efficient Multivariate Dynamic Screening System Jun Li a and Peihua Qiu b a Department of Statistics, University of California at Riverside b Department of Biostatistics, University

More information

Monitoring Multivariate Data via KNN Learning

Monitoring Multivariate Data via KNN Learning Monitoring Multivariate Data via KNN Learning Chi Zhang 1, Yajun Mei 2, Fugee Tsung 1 1 Department of Industrial Engineering and Logistics Management, Hong Kong University of Science and Technology, Clear

More information

Correction factors for Shewhart and control charts to achieve desired unconditional ARL

Correction factors for Shewhart and control charts to achieve desired unconditional ARL International Journal of Production Research ISSN: 0020-7543 (Print) 1366-588X (Online) Journal homepage: http://www.tandfonline.com/loi/tprs20 Correction factors for Shewhart and control charts to achieve

More information

Likelihood Ratio-Based Distribution-Free EWMA Control Charts

Likelihood Ratio-Based Distribution-Free EWMA Control Charts Likelihood Ratio-Based Distribution-Free EWMA Control Charts CHANGLIANG ZOU Nankai University, Tianjin, China FUGEE TSUNG Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong

More information

An optimization model for designing acceptance sampling plan based on cumulative count of conforming run length using minimum angle method

An optimization model for designing acceptance sampling plan based on cumulative count of conforming run length using minimum angle method Hacettepe Journal of Mathematics and Statistics Volume 44 (5) (2015), 1271 1281 An optimization model for designing acceptance sampling plan based on cumulative count of conforming run length using minimum

More information

TESTS FOR EQUIVALENCE BASED ON ODDS RATIO FOR MATCHED-PAIR DESIGN

TESTS FOR EQUIVALENCE BASED ON ODDS RATIO FOR MATCHED-PAIR DESIGN Journal of Biopharmaceutical Statistics, 15: 889 901, 2005 Copyright Taylor & Francis, Inc. ISSN: 1054-3406 print/1520-5711 online DOI: 10.1080/10543400500265561 TESTS FOR EQUIVALENCE BASED ON ODDS RATIO

More information

D Chart: An Efficient Alternative to Monitor Process Dispersion

D Chart: An Efficient Alternative to Monitor Process Dispersion Proceedings of the World Congress on Engineering and Computer cience 2011 Vol II WCEC 2011, October 19-21, 2011, an Francisco, UA Chart: An Efficient Alternative to Monitor Process ispersion addam Akber

More information

An exponentially weighted moving average scheme with variable sampling intervals for monitoring linear profiles

An exponentially weighted moving average scheme with variable sampling intervals for monitoring linear profiles An exponentially weighted moving average scheme with variable sampling intervals for monitoring linear profiles Zhonghua Li, Zhaojun Wang LPMC and Department of Statistics, School of Mathematical Sciences,

More information

Cusum charts for preliminary analysis of individual observations

Cusum charts for preliminary analysis of individual observations Cusum charts for preliminary analysis of individual observations Alex J. Koning æ Econometric Institute Erasmus University Rotterdam P.O. Box 1738 NL-3000 DR Rotterdam koning@few.eur.nl Ronald J.M.M. Does

More information

Run sum control charts for the monitoring of process variability

Run sum control charts for the monitoring of process variability Quality Technology & Quantitative Management ISSN: (Print) 1684-3703 (Online) Journal homepage: http://www.tandfonline.com/loi/ttqm20 Run sum control charts for the monitoring of process variability Athanasios

More information

Multivariate Process Control Chart for Controlling the False Discovery Rate

Multivariate Process Control Chart for Controlling the False Discovery Rate Industrial Engineering & Management Systems Vol, No 4, December 0, pp.385-389 ISSN 598-748 EISSN 34-6473 http://dx.doi.org/0.73/iems.0..4.385 0 KIIE Multivariate Process Control Chart for Controlling e

More information

A New Bootstrap Based Algorithm for Hotelling s T2 Multivariate Control Chart

A New Bootstrap Based Algorithm for Hotelling s T2 Multivariate Control Chart Journal of Sciences, Islamic Republic of Iran 7(3): 69-78 (16) University of Tehran, ISSN 16-14 http://jsciences.ut.ac.ir A New Bootstrap Based Algorithm for Hotelling s T Multivariate Control Chart A.

More information

Confirmation Sample Control Charts

Confirmation Sample Control Charts Confirmation Sample Control Charts Stefan H. Steiner Dept. of Statistics and Actuarial Sciences University of Waterloo Waterloo, NL 3G1 Canada Control charts such as X and R charts are widely used in industry

More information

A New Model-Free CuSum Procedure for Autocorrelated Processes

A New Model-Free CuSum Procedure for Autocorrelated Processes A New Model-Free CuSum Procedure for Autocorrelated Processes Seong-Hee Kim, Christos Alexopoulos, David Goldsman, and Kwok-Leung Tsui School of Industrial and Systems Engineering Georgia Institute of

More information

MCUSUM CONTROL CHART PROCEDURE: MONITORING THE PROCESS MEAN WITH APPLICATION

MCUSUM CONTROL CHART PROCEDURE: MONITORING THE PROCESS MEAN WITH APPLICATION Journal of Statistics: Advances in Theory and Applications Volume 6, Number, 206, Pages 05-32 Available at http://scientificadvances.co.in DOI: http://dx.doi.org/0.8642/jsata_700272 MCUSUM CONTROL CHART

More information

ARL-unbiased geometric control charts for high-yield processes

ARL-unbiased geometric control charts for high-yield processes ARL-unbiased geometric control charts for high-yield processes Manuel Cabral Morais Department of Mathematics & CEMAT IST, ULisboa, Portugal IST Lisboa, March 13, 2018 Agenda 1 Warm up A few historical

More information

An-Najah National University Faculty of Engineering Industrial Engineering Department. Course : Quantitative Methods (65211)

An-Najah National University Faculty of Engineering Industrial Engineering Department. Course : Quantitative Methods (65211) An-Najah National University Faculty of Engineering Industrial Engineering Department Course : Quantitative Methods (65211) Instructor: Eng. Tamer Haddad 2 nd Semester 2009/2010 Chapter 3 Discrete Random

More information

The Effect of Level of Significance (α) on the Performance of Hotelling-T 2 Control Chart

The Effect of Level of Significance (α) on the Performance of Hotelling-T 2 Control Chart The Effect of Level of Significance (α) on the Performance of Hotelling-T 2 Control Chart Obafemi, O. S. 1 Department of Mathematics and Statistics, Federal Polytechnic, Ado-Ekiti, Ekiti State, Nigeria

More information

Design and Application of Risk Adjusted Cumulative Sum (RACUSUM) for Online Strength Monitoring of Ready Mixed Concrete

Design and Application of Risk Adjusted Cumulative Sum (RACUSUM) for Online Strength Monitoring of Ready Mixed Concrete INDIAN INSTITUTE OF MANAGEMENT AHMEDABAD INDIA Design and Application of Risk Adjusted Cumulative Sum (RACUSUM) for Online Strength Monitoring of Ready Mixed Concrete Debasis Sarkar Goutam Dutta W.P. No.2008-08-01

More information