Applying Genetic Algorithms to Solve the Fuzzy Optimal Profit Problem

Size: px
Start display at page:

Download "Applying Genetic Algorithms to Solve the Fuzzy Optimal Profit Problem"

Transcription

1 JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 8, () Applying Genetic Algoriths to Solve the Fuzzy Optial Profit Proble FENG-TSE LIN AND JING-SHING YAO Departent of Applied Matheatics Chinese Culture University Taipei, Taiwan E-ail: Departent of Matheatics National Taiwan University Taipei, 6 Taiwan This study investigated the application of genetic algoriths for solving a fuzzy optiization proble that arises in business and econoics. In this proble, a fuzzy price is deterined using a linear or a quadratic fuzzy deand function as well as a linear cost function. The obective is to find the optial fuzzy profit, which is derived fro the fuzzy price and the fuzzy cost. The traditional ethods for solving this proble are () using the etension principle, and () using the interval arithetic and α-cuts. However, we argue that the traditional ethods for solving this proble are too restrictive to produce an optial solution, and that an alternative approach is possibly needed. We use genetic algoriths to obtain an approiate solution for this fuzzy optial profit proble without using the ebership functions. We not only give epirical eaples to show the effectiveness of this approach, but also give theoretical proofs to validate the correctness of the algorith. We conclude that genetic algoriths can produce good approiate solutions when applied to solve fuzzy optiization probles. Keywords: genetic algoriths, fuzzy sets, fuzzy nubers, fuzzy optiization profit proble, fuzzy deand. INTRODUCTION In this study, we investigated the application of genetic algoriths to solve a fuzzy optiization proble that arises in business and econoics. This proble is often involved in the study of how changes in such variables as production or price will affect other variables such as revenue or profit [4,, ]. One proble instance is the fuzzy optial profit proble [], which is briefly depicted as follows. In a onopolist arket, producers can control arket prices and product quantities. The deand for a certain coodity is related to its price by a deand function P (. This eans that, as the price increases, deand usually falls, and that as the price falls, deand rises. Since revenue is equal to the price per unit ties the quantity sold, we can deterine the revenue received for selling units of the coodity as R ( P(. The cost of producing units of a certain coodity is given by a cost function C ( and the basic relation between profit, revenue and cost is forulated as N ( R( C(. The onopolist can thus easily obtain the aiu profit by doing soe siple calculations. Received August 9, ; revised October 5, ; accepted January 9,. Counicated by Chuen-Tsai Sun. 563

2 564 FENG-TSE LIN AND JING-SHING YAO On the other hand, in a perfect copetitive arket, the deand is no longer a fied value, even for the sae price function P (. The price, of course, will fluctuate at any tie in order to reflect various conditions in the arket. To deal with this kind of iprecise data, fuzzy sets provide a powerful tool for odeling and solving the optiization proble. Thus, this leads to the use of the fuzzy nuber to represent the fuzzy deand and the fuzzy profit. The optial price can then be obtained using the ebership functions and the centroid of the fuzzy profit. Nevertheless, the ain issue in this proble is as follows. Consider a quadratic algebraic epression a + b for a, b real paraeters and where is a real variable. Let y N(a, b, a + b. After substituting the triangular fuzzy nubers A, B, and into a + b for a, b, and, respectively, we obtain the fuzzy set A B +. The triangular fuzzy nubers are represented by A ( a, a, a + ), B ( b w, b, b + w ), and ( δ,, + δ ) together with their ebership functions µ(a A ), µ(b B ), and µ( ), respectively. Accordingly, there are two aor traditional ethods for evaluating the above fuzzy epression in the literature [, 3]. The first ethod for finding the value of A + B involves using the etension principle. Let Π(a, b, denote the iniu of µ(a A ), µ(b B ), and µ( ). Assue that we set Y equal to A + B. Then the ebership function for Y is defined by µ (y Y ) sup{π(a, b, N(a, b, y}. To find α-cuts of Y, let Φ(α) {N(a, b, a A ( α ), b B ( α ), (α)}, α. Then we have Φ(α) Y ( α ). Alternatively, the second ethod involves evaluating A + B using interval arithetic and α-cuts. For any triangular fuzzy nubers A and B, the following operations hold: () ( A B )(α) A ( α ) B ( α), and () ( A ± B )(α) A ( α ) ± B ( α ). A B and A ± B are coputed using the etension principle, and A ( α ) B ( α), A ( α ) ± B ( α ) are found using interval arithetic. Let Z ( α ) A ( α ) ( α) ( α) + B ( α ) ( α), for α. We can see that Y ( α ) is a subset of Z ( α ). Obviously, evaluating a fuzzy algebraic epression using interval arithetic and α-cuts can produce a larger fuzzy set than can evaluating using the etension principle. In addition, these traditional ethods for solving fuzzy equations are very often too restrictive to produce a solution [3]. The other disadvantages, as noted in [7], are that solving fuzzy equations is very coplicated due to the lack of inverse operators, and that the ultiple occurrence of paraeters in an epression results in increased iprecision. In this study, we applied a genetic algorith to obtain an approiate solution to the fuzzy optial profit optiization proble. When genetic algoriths are applied to solve this proble, the fuzzy equation coputation requires no etension principle or interval arithetic and α-cuts. The genetic algorith uses only the usual evolution. Assue that a is an arbitrary fuzzy set on the interval [, ], and that N ( ) is a b fuzzy profit. In the genetic algorith approach, we do not need to define the a ebership functions of N ( ). Instead, the interval [, ] is equally divided into b a partitions. Let,,,, be the partition points. Let ( ) [,] b µ,

3 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM 565,,,, be the ebership grade of in. Thus, we obtain a discrete fuzzy set ( µ, µ,..., µ ), where µ,,,,, is a rando nuber in [, ]. In other words, we wish to find a in [, ] by aiizing b N ( ) via the genetic algoriths. However, we cannot aiize N ( ) directly since it is a fuzzy set. Instead, we can copute the centroid of the fuzzy profit for the aiization proble. The centroid can be used as the fitness value for the evolution of the genetic algoriths. Therefore, the obective is to siply find a vector in [, ] + that aiizes the centroid. This paper is organized as follows. Section deals with the optial price for fuzzy profit based on the linear and quadratic deand functions. First, we discuss the traditional ethod for obtaining the optial solution using fuzzy optiization and then introduce proble discretization for the purpose of the genetic algorith application to obtain an approiate solution to this proble. In section 3, we design a genetic algorith for solving the fuzzy optial profit proble. Section 4 gives two illustrative eaples. Section 5 discusses the ain results of this work. Finally, we state our conclusions in section 6.. Linear Deand Function. OPTIMAL PRICE FOR FUZZY PROFIT First, a linear deand function is introduced. Assue that the linear function is P ( a b, a / b () and that the cost function is C ( e + g + k, () where a, b, e, g, k are known positive nubers, b ( a g) < a( b + k), and g < a. The profit function is defined by P( C( e + ( a g) ( b k), a / b N ( +. (3) In a onopolist arket, the profit function is surely (3). Since N ( a g ( b + k), a g we have ( ). According to the assuption that b ( a g) < a( b + k), we ( b + k) have a (, ]. Hence, the aiu profit is b N( a g ( a g) ) N e +. (4) ( b + k) 4( b + k) a P( However, in a perfect copetitive arket, the deand will vary, even for b the sae price P (. As a result, the deand can be represented by the fuzzy nuber.

4 566 FENG-TSE LIN AND JING-SHING YAO. Quadratic Deand Function Net, consider the quadratic deand function. Given the deand function a b + c,, (5) P( and the cost function C ( e + g + k,, (6) where all variables a, b, c, d, e, g, and k are known positive nubers, the relations of this quadratic function are b b 4ac () If b 4ac, then set. c b () If b 4ac <, then set c. Fro (5) and (6), we have the profit function as follows: 3 N ( e + ( a g) ( b + k) + c, Consider a onopolist arket.. (7) By differentiating (7), we obtain the following equation: N ( ( a g) ( b + k) + 3c,. The aiu is found to be 3c ( b + k) + ( a g). (8) The discriinant for (8) is D ( b + k) 3c( a ). g Case. If D >, then (8) has two roots, b + k D d and b + k + D d. 3c 3c We obtain N 3c( d )( ). Then, we have the following three possibilities: ( d () When < d < d <, a N( a( N( d), N( )). (9) () When < d < < d, a N( N( d). () (3) When < d < d, a N( N( ). () b + k Case. If D, then N ( 3c( ) > and a N( N( ) 3c. () Case 3. If D <, then N ( > < < and a N( N( ). (3) Assue that the optial deand is. Then the aiu profit is N ( ). Note that in a onopolist arket, the deand is uniquely deterined by the given price P(. However, in a perfect copetitive arket, the deand and the price do not necessary a P( depend on the deand function. Therefore, the deand b will vary, even for the sae price P(. This is a copletely different interpretation of the deand in a copetitive arket copared with that in a onopolist arket. As a result, the deand

5 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM 567 should be represented using linguistic variables, such as if the price is P(, then the deand is in the vicinity of, to obtain a reasonable description..3 Traditional Fuzzy Methods Obviously, the deand in a copetitive arket can be represented using the triangular fuzzy nuber. Fro (7), we obtain the fuzzy profit N ( ) e + (a g) (b + k). Let N ( y. Then (3) becoes (b + k) (a g) + e + y. (4) Fro (4), we can see that D ( y) ( a g) 4( b + k)( e + y), if y N( ). If D ( y) >, then (4) has two roots, which are quadratic forulas, shown as follows: ( a g D( y) r y), ( b + k) a g + D( y) r( y). (5) ( b + k) Otherwise, if D(y), (4) has a duplicate root, then a g r( y) r ( y). (6) ( b + k) After applying the etension principle ethod, we obtain the ebership function of the fuzzy profit N ( ) as N( ( y) y a[ ( r ( )), ( ( ))], ( sup ( ) y r y if y N N (, otherwise ), (7) where the ebership function of the triangular fuzzy nuber is ( ) and the ebership function of the fuzzy profit N ( ) is N ( )( y). Once we have N ( )( y), the centroid of N ( ), E(, can be obtained using the following equation: E( N ( ) N ( ) yn( ( y) dy. (8) N( ( y) dy Then the best solution of the fuzzy optial profit proble is obtained. Siilarly, consider the quadratic function. After we apply the etension principle, the ebership function of N ( ) is µ N ( ) ( y) sup µ ( if N - (y). Note N ( y) that N( y is a cubic equation. Clearly, producing N ( y ) for a cubic equation is not an easy task. (see section 5.3. for a detailed discussion)..4 Genetic Algoriths Approach The genetic algoriths approach is depicted as follows. Let the triangular fuzzy nuber (,, + ) be replaced by an arbitrary fuzzy set in an interval [, ], where is defined in (5) (see Fig. ).

6 568 FENG-TSE LIN AND JING-SHING YAO µ3 µ µ µ ( ) Fig.. Fuzzy set. The interval [, ] is divided into partitions,, where,,, are the partition points. Let the ebership grade of at be ( ) µ,,,, and µ [,]. Then, a discrete fuzzy set is obtained as follows: ( µ, µ,..., µ ) µ µ µ (9) Fro N ( y, if each N ( k ), k,,,, is different, then the fuzzy profit is defined by µ µ µ N ( ) N ( µ, µ,..., µ ) , () N ( ) N ( ) N ( ) and the centroid is defined by N N ( ) µ θ ( ( )). () µ Fro (7), let N ( y ; we obtain a cubic equation: 3 c ( b + k) + ( a g) ( e + y). Assue that there are three roots for this equation. Let N( ) N( ) N( ) y, 3 ( µ, µ,..., µ N )( y where 3. According to the etension principle, we obtain N ) sup ( a( ( ), ( ), ( )) a[,, ] y N ( µ 3 µ µ 3. Eq. () becoes a[ µ, µ ], µ. The centroid of ) Q 3 N ( is θ ( N ( )), () N( ) P where P µ + a( µ, µ, µ 3),, 3 and Q N( ) µ + N( )(a[ µ, µ, µ 3]).,, 3 Let the centroid of the fuzzy profit N ) N( µ, µ,..., µ ) be θ N ( )) centroidof N( µ, µ,..., µ ). ( (

7 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM 569 Finally, a vector is deterined in [, ] + (real nubers) in a population that aiizes )) by eans of genetic algoriths. 3. DESIGNING A GENETIC ALGORITHM FOR SOLVING THE FUZZY OPTIMAL PROFIT PROBLEM Genetic algoriths are stochastic search techniques based on the principles and echaniss of natural genetics and selection [8, 5]. Genetic algoriths grew out of Holland s [] study on adaptation in artificial and natural systes. The basic concept of genetic algoriths is that they start with a population of randoly generated candidates and evolve towards better solutions by applying genetic operators, such as crossover and utation, odeled on natural genetic inheritance and Darwinian survival-of-the-fittest [8, 5]. Therefore, genetic algoriths are theoretically and epirically have been proven to process robust search capabilities in cople spaces, thus offering a valid approach to probles requiring efficient and effective searching [6]. Genetic algoriths can be used to copute the ebership functions of fuzzy sets [, 3]. Given soe functional apping for a syste, soe ebership functions and their shapes are assued for the various fuzzy variables defined for a proble [6]. The ebership functions are coded as bit strings that are then concatenated. An evaluation function is used to evaluate the fitness of each set of ebership functions. There are two possible ways to integrate fuzzy logic and genetic algoriths [9]. One involves the application of genetic algoriths for solving optiization and search probles related to fuzzy systes [,, 7]. The other, is the use of fuzzy tools and fuzzy logic-based techniques for odeling different genetic algorith coponents and adapting genetic algorith control paraeters, with the goal of iproving perforance [9, 4, 8, 9]. Now, a genetic algorith for solving the fuzzy optial profit proble is given below. Step. Generate an initial population. An initial population of size n is randoly generated fro [, ] + according to the unifor distribution in the closed interval [,]. Let the population be ( µ, µ,..., ) µ µ µ µ , where,,, n and µ i is a real nuber in [, ], i,,,,. Each individual in a population is a chroosoe. Step. Calculate the fitness value for each chroosoe. For each chroosoe,,,, n, the centroid )) is calculated as the fitness value. The chroosoes in the population can be rated in ters of their fitness n values. Let the total fitness value of the population be T θ ( N ( )). The cuulative k fitness value (partial su) for each chroosoe, S θ ( N ( )), k,,, n, is k calculated. Net, intervals I [, S ], I [S -, S ],, 3,, n, and I n [S n-, S n ]

8 57 FENG-TSE LIN AND JING-SHING YAO are constructed for the purpose of selection. In our ipleentation, a roulette wheel approach [6, 8] is adopted as the selection echanis. Step 3. Selection and reproduction. Reproduction is a process in which each chroosoe is copied according to the selection process. The selection process begins with spinning of the roulette wheel n ties. Each tie, a single chroosoe is selected fro the current generation to create a new generation. The selection process is as follows. Each tie, a rando nuber r fro the range [, T] is generated. If r I, then chroosoe is selected; otherwise, the kth chroosoe k, k n, is selected if r I k. This selection process is continued until the new population has been created. Finally, the new population is renaed Y, Y, Y3, in the order they were picked. The probability of selection for each chroosoe )) such that it appears in the new population is, and its )) epected value is n T )) to go on into the net population. T. Thus, this procedure tends to choose ore with higher Step 4. Perfor crossover. Crossover is the key to genetic algoriths power. The purpose of crossover is to generate rearrangeents of coadapted groups of inforation fro high perforance structures [8]. The crossover ethod used here is the one-point ethod, which randoly selects one cut-point and echanges the right parts of two parents to generate offspring. Each pair ( Y, Y ), ( Y 3, Y 4 ),, ( Y n, Y n ), where n is even, produces two children via crossover. Let p be the probability of a crossover, p. Usually, p is between.6 and.9 [8], so we epect that, on average, 6% to 9% of the chroosoes will undergo crossover. Consider two chroosoes, Y and Y, for crossover. A rando nuber r is generated in the interval [, ]. If r p, then crossover is perfored on Y and Y. Otherwise, if r > p, the two children Y ' and Y ' are identical to their parents, Y and Y. Suppose that crossover needs to be perfored. Another rando integer nuber u is generated fro the range [, ]. Assue that u equals 3; the two chroosoes Y and Y are cut after the fourth position, and the ' offspring Y and Y ' are generated by echanging the right parts of each chroosoe. Step 5. Perfor utation. Mutation is a background operator that produces rando changes in various chroosoes [6]. Mutation resets one randoly selected position to a real nuber in [, ] or to zero with a probability equal to the utation rate (see Eaples in section 4). Let q be the probability of a utation, q. Usually q is a very sall value, around. to. [8], so we epect that, on average,.% to % of the total population will undergo utation. There are n positions in the whole population. We epect. n to. n utations per generation. For eaple, consider chroosoe ' Y. Let w i be a rando nuber in [, ], i. If w i < q, then the i+th position ' in Y is reset to a real rando nuber in [, ] or to zero. After the utation is copleted on the whole population, we let Y,,,, n. One iteration of the

9 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM 57 genetic algorith has been copleted. Step through step 5 is done K ties, where K is the aiu nuber of iterations. Finally, the algorith is terinated after K generations are produced. Let the last population be,,..., n. The centroid of each chroosoe, n, in the last population ust now be calculated. The aiu value of k )) is the best chroosoe in the population, i.e., a )) )) for soe k {, n,..., n}. The best chroosoe could be a sub-optial or the optial solution for the proble. Let the best chroosoe be (,,..., µ k k k k µ k µ k µ k ) µ µ If P ( i ) a bi + ci, for each i, where i,,,, are different values, then the fuzzy price function is P( k µ k k ) a b k + c k a b + c a b + c Therefore, the centroid of P ( ) is This is an estiate of the aial profit in the fuzzy sense. k k k µ µ +. (3) a b + c ( a b + c ) µ k E( P( k )). (4) µ k In particular, if P( ) P( ), i, then i,, µ k i µ k + of (3) will becoe P( ) P( ) i a( µ k, i, µ P( ) i k, ). In the following, two properties are given to show how good the final solution obtained by genetic algoriths is, and it is copared with the known crisp optial value. Property. The optial value k )) obtained using genetic algoriths is less than or equal to the crisp optial value N ( ). That is, )) N ( ). Proof. For any discrete fuzzy set, ( µ, µ,..., µ ) µ µ µ [, ], µ [,],,,, its fuzzy profit is N ( ) ) k , where N ( µ, µ,..., µ. When N ( k ), k,,,, are different values, the centroid is obtained by N N ( ) µ θ ( ( )). Note that N ( ) is the crisp optial solution (i.e. aiu µ profit). For each [, ],,,,, we have N ( ) N( ),.

10 57 FENG-TSE LIN AND JING-SHING YAO Since µ and, we obtain N( ) µ N( ) µ. Therefore, )) N( ). Furtherore, for each N ( ),,,,, if there eist three identical values, e.g. N ( ) N ( ) N ( 3), 3, then fro (), the centroid of Q N ( ) is θ ( N ( )), where P µ + a( µ, µ, µ 3) and P,, 3 Q N( ) µ + N( )(a[ µ, µ, µ 3]). We can see that )) N( ).,, 3 Therefore, we obtain k )) N( ). Q.E.D Property. The search space for genetic algoriths includes the crisp optial profit solution. However, the probability that the value of )) is equal to the crisp value N ( ) is sall. Proof. The crisp optial deand is in the interval [, ]. When the interval [, ] is divided into partitions,,,,,, we have the possibility that one of b the partition includes. Let i be, i {,,, }. Fro (9), we have µ µ µ , where µ is a rando nuber in [, ],,,,. We can see that there is a possibility of having µ for i and µ. Thus, we have µ i i i µ. Fro (), we obtain N,..., µ,,..., ) ( i i µ i N ) µ i. Then, the centroid θ ( N (,...,, µ,,..., ) i N ( i ) is equal to N ( ). N( ) Obviously, the genetic algorith search space theoretically includes the crisp optial profit solution. However, the probability that the value of )) equals the crisp value N ( ) is sall. Q.E.D. ( i 4. EMPIRICAL RESULTS Two epirical eaples are given below to illustrate the effectiveness of genetic algoriths when we applying the to solve the fuzzy optial profit proble. Eaple. The first instance is as follows: the price function, P(, 5; the cost function, C( + ; and the profit function, N( P( C( + 99, 5. Since N ( 99 4, we obtain Thus, the optial solutions for the crisp case are P(4.75) 5.5 and N(4.75) 5.5. Now, 5 consider that the deand is vague. Let. We have k k 5k, k,,,, which eans that ten partition points are given. Then, we obtain N( ) -, N( )

11 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM , N( ) 78, N( 3 ) 5, N( 4 ) 7, N( 5 ) 5, N( 6 ) 6, N( 7 ) 5, N( 8 ) 75, N( 9 ) 395, and N( ) 6. Fro () and (), we have N( ) µ N ( ) N( µ, µ,..., µ ) and the centroid )). The estiated price µ in the fuzzy sense is given by E( P( )) P( µ ) µ. The paraeters for genetic algoriths are () the probability of a crossover p.8 and () the probability of a utation q.3. An essential feature of genetic algoriths when they are used to solve the fuzzy optial profit proble is their effectiveness in solving the fuzzy proble. Two different approaches were taken in this study for the purpose of analyzing the effectiveness of the algorith. The first approach was based on using the sae nuber of generations but with different population sizes. The second approach was based on using the sae population size but with different nubers of generations. Each approach had five or si runs. In the first approach, the nuber of generations was, and the population size for each run was,, 3, 4, and 5, respectively. The best results for obtained in five runs are shown in Fig.. The partition points are 5,, 5,, 5,, etc. We can see that the genetic algorith was trying to approiate ( ). for, 5, and 3, and ( ). otherwise, where a bell-shaped discrete fuzzy set was obtained. In these runs, one of the best solutions obtained was (.4,.,.3,.,.9,.98,.87,.6,.4,.3,.) with the largest value being )) and the estiated price in the fuzzy sense being E ( P( )) A coparison of the above obtained result with that of the crisp case is % and % ebership partition points run run run 3 run 4 run 5 Fig.. A bell-shaped discrete fuzzy set obtained using the first approach. The nuber of generations is, and the population size in each run is,, 3, 4, and 5, respectively. As for the second approach, the population size was in all si runs. The nuber of generations in each run was,,,, 4,, 6,, 8,, and

12 574 FENG-TSE LIN AND JING-SHING YAO,, respectively. The best results for obtained in si runs are shown in Fig. 3. Once again, the genetic algorith was trying to approiate ( ). for, 5, and 3, and ( ). otherwise, in these runs. Note that the curve for run 6 shown in Fig. 3 has a bell-shaped discrete fuzzy set with ().86, (5).9, and (3).58. ebership partition points 5 run run run 3 run 4 run 5 run 6 Fig. 3. A bell-shaped discrete fuzzy set was obtained fro run 6, which had a population size of and, generations. If we let the utation reset one randoly selected position to zero (i.e. force the ebership grade to ), the best result obtained in each run becae (.,.,.,.,.,.8,.,,.). In this case, the aiu profit obtained by genetic algoriths was the discrete fuzzy set. 8 (where the ebership grade k 5 was.8 when was 5) with an optial value of 5 and an estiated price of 5. 5 Since the partition points k k 5k, k,,, did not include the optial deand 4.75 (the nearest point was 5 5.), the best profit obtained by the genetic algoriths was only 5 (the optial profit in the crisp case was 5.5). Finally, consider the situation when ore partition points are given. Let 5. The nuber of partition points was 5. Fig. 4 shows the ebership curves obtained fro si runs based on using the sae population size but different nubers of generations. The nuber of generations in each run was,,,, 4,, 6,, 8,, and,, respectively. All these curves are trying to approiate ( ). for 3.3 and 6.7. Note that the curve fro run 6 has a bell-shaped discrete fuzzy set with ().64, (3.33).86, (6.67).85, and (3).76. ebership partition points run run run 3 run 4 run 5 run 6 Fig. 4. The bell-shaped discrete fuzzy set obtained fro run 6 using the second approach with 5 partition points.

13 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM 575 Eaple. The proble instance is as follows: the price function is a quadratic, P( +, 5; the cost function is C( +, ; and the profit function is N( + + 3, 5. Since b 4ac 44 <, we have / 5 and N ( + ( 5) + 5 >. In the crisp case of the proble, when 5, the optial solutions are P(5) 76 and N(5) 365. Now, 5 consider that is vague. Let ; we have k k.5k, k,,,. We obtain N( ) -, N( ) 37.65, N( ) 8, N( 3 ).875, N( 4 ) 58, N( 5 ) 93.5, N( 6 ) 7, N( 7 ) 6.375, N( 8 ) 94, N( 9 ) 38.65, and N( ) 365. The centroid is E( P( )) P( µ )) N( µ ) µ, and the optial price in the fuzzy sense is ) µ. The paraeters and the two approaches used for genetic algoriths were all the sae as in Eaple. The best results for obtained in five runs using the first approach with the sae nuber of generations but with different population sizes are shown in Fig. 5. The partition points are.5,.,.5,.,.5,, etc. We can see that the genetic algorith was trying to approiate ( ). for 4.5, 5. and ( ). otherwise, where a half bell-shaped discrete fuzzy set was obtained. In these runs, one of the best solutions obtained was (.,.4,.7,.58,.38,.5,.6,.,.7,.849,.98) with the largest value being )) and the estiated price in the fuzzy sense being E ( P( )) A coparison of the results obtained with that of the crisp case is %.543%, and %.63% ebership partition points run run run 3 run 4 run 5 Fig. 5. A half bell-shaped discrete fuzzy set obtained using the first approach. For the second approach, using the sae population size but with different nubers of generations, the best results for obtained fro si runs are shown in Fig. 6.

14 576 FENG-TSE LIN AND JING-SHING YAO Fro Fig. 6, we can see that the genetic algorith was trying to approiate ( ). for 4.5, 5. and ( ). otherwise, where a half bell-shaped discrete fuzzy set was again obtained. Once again, if we let the utation reset one randoly selected position to zero, the best result obtained in each run becae (.,.,.,,.,.,.95). In this case, the aiu profit obtained by the.95 genetic algoriths was the discrete fuzzy set k 5. (the ebership grade was.95 when was 5.) with a value of ebership run run run 3 run 4 run 5 run 6 partition points Fig. 6. A half bell-shaped discrete fuzzy set obtained using the second approach. 5. DISCUSSION In this section, we will point out that our work has produced the following ain results based on the application of genetic algoriths to solve the fuzzy optial profit proble. 5. Genetic Algoriths Can Find the Optial Solution As eplained in Section, the crisp optial deand, was in (, ], and we can take as a suitable value to divide the interval (, ] into sall partitions, i.e.,,,,. According to Properties and, let i for soe i b {,,,, }. Fro (9), the rando nubers µ, µ,..., µ, generated fro the interval [, ], should have a possibility of µ, i and µ. Therefore, (9) becoes,...,, µ,,...,) ( i µ i µ i i. i i µ According to (), we have N (,...,, µ i,,..., ), where N ( ) is the crisp aiu profit. We conclude N( ) that the search space for genetic algoriths includes the crisp optial solution and crisp near-optial solutions.

15 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM Mutation Strategies In our ipleentation, two strategies are used to perfor utation in genetic algoriths. Norally, we use Strategy. Strategy resets one randoly selected position to a real nuber in [, ] (sets the ebership grade to the other value). On the other hand, Strategy resets one randoly selected position in a chroosoe to zero (forces the ebership grade to zero). Eaple in Section 4 shows that Strategy had in ore likely than Strategy to include the crisp optial solution in its search space. Epirical results show that the speed of convergence to the crisp optial solution using Strategy was uch faster than that achieved using Strategy. The underlying theore for Strategy is based on Property. 5.3 Coparison Between the Traditional Fuzzy Method and Genetic Algoriths The aor difference in obtaining the fuzzy optial profit between using the traditional fuzzy ethod [, ] and genetic algoriths is stated as follows Using the etension principle ethod As eplained in Section, when the deand is fuzzified into a triangular fuzzy nuber (,, + ), < <, >, obviously, the fuzzy profit is N ( ). By applying the etension principle ethod, we can obtain the ebership function of N ( ), µ N ( ) ( y) sup µ (. Since N( y is a cubic function, i.e., e + (a N ( y) g) (b + k) + c 3, it is difficult to obtain N ( y ). As a result, the ebership function µ N ( ) ( y) is difficult to obtain by applying the etension principle. Furtherore, if we let the three roots of N( y be t (y), t (y), and t 3 (y), then we can see that µ N ( ) ( y) sup in[ µ ( t( y )), µ ( t ( y)), µ ( t 3 ( y)) ], where t +, t + t t), t +, otherwise µ ( Obviously, it is also difficult to obtain µ N ( ) ( y) fro the above equation. In this case, the proposed genetic algorith approach can be considered as an alternative way to find a near optial solution Using genetic algoriths With the genetic algorith approach, however, no specific fuzzy set is needed (i.e. use of a triangular fuzzy nuber or other special fuzzy sets is not necessary). Initially, a

16 578 FENG-TSE LIN AND JING-SHING YAO population of discrete fuzzy sets, µ + µ + + µ,,,, n, is created based on the generation of rando nubers in [, ]. After the copletion of the evolution by genetic algoriths, the best discrete fuzzy set in the population is k µ + k µ + + µ k. The fuzzy profit is N ( ) k µ k + N( ) k µ k + N( ) + µ k. After the centroid ethod is used for defuzzification, the estiated profit in N( ) the fuzzy sense is obtained: )) k N( ) µ k µ k. Consider Eaple in section 4. The crisp optial solutions are P(5) 76 and N(5) 365. The best discrete fuzzy set obtained fro si runs is given in Fig. 6, represented by µ k. The fuzzy µ profit is N ( k ) N( N( µ ) µ ), N( ),,,, and the centroid is )) In addition, the best discrete fuzzy set in the population when using Strategy is and the centroid is )) k k.95. The fuzzy profit is N ( 5. ) k N(5.) k.95, CONCLUSIONS This study has investigated the genetic algorith approach to solving fuzzy optiization equations without using the ebership functions of fuzzy nubers. When genetic algoriths are applied to solve fuzzy equations, the coputation uses neither the etension principle nor the interval arithetic and α-cuts. The results of this study ay lead to the developent of effective genetic algoriths for solving general fuzzy optiization probles. In suary, the fuzzy concept of the proposed genetic algorith approach is different and gives alost the sae results as the traditional ethods. ACKNOWLEDGEMENT The authors are grateful to the anonyous referees, whose valuable coents helped to iprove the content of this paper.

17 APPLYING GENETIC ALGORITHMS TO SOLVE THE FUZZY OPTIMAL PROFIT PROBLEM 579 REFERENCES. J. J. Buckly and Y. Hayashi, Fuzzy genetic algorith and applications, Fuzzy Sets and Systes, Vol. 6, 994, pp J. J. Buckley and Y. Qu, On using α-cuts to evaluate fuzzy equations, Fuzzy Sets and Systes, Vol. 38, 99, pp J. J. Buckley and Y. Qu, Solving linear and quadratic fuzzy equations, Fuzzy Sets and Systes, Vol. 38, 99, pp J. J. Buckley, Solving fuzzy equations in econoics and finance, Fuzzy Sets and Systes, Vol. 48, 99, pp L. Davis, ed., Genetic Algoriths and Siulated Annealing, San Mateo, CA: Morgan Kaufann, M. Gen and R. Cheng, Genetic Algoriths & Engineering Design, John Wiley & Sons, Inc., New York, R. E. Giachetti, Evaluating engineering functions with iprecise quantities, 7 th International Fuzzy Systes Association Congress, Prague, Czech Republic, D. E. Goldberg, Genetic Algoriths in Search, Optiization, and Machine Learning, Addison-Wesley, Reading, MA, F. Herrera and M. Lozano, Fuzzy genetic algoriths: issues and odels, Dept. of Science and A. I., University of Granada, Technical Report No. 87, Granada, Spain, F. Herrera, M. Lozano, and J. L. Verdegay, Applying genetic algoriths in fuzzy optiization probles, Fuzzy Sets & Artificial Intelligence, Vol. 3, 994, pp J. Holland, Adaptation in Natural and Artificial Systes, University of Michigan Press, Ann Arbor, A. Hoaifar and E. McCorick, Siultaneous design of ebership functions and rule sets for fuzzy controllers using genetic algoriths, IEEE Transcations on Fuzzy Systes, Vol. 3, 995, pp C. L. Karr and E. J. Gentry, Fuzzy control of ph using genetic algoriths, IEEE Transactions of Fuzzy Systes, Vol., 993, pp M. A. Lee and H. Takagi, Dynaic control of genetic algoriths using fuzzy logic techniques, in Proceedings of Fifth International Conference on Genetic Algoriths (ICGA 93), 993, pp M. Mitchell, An Introduction to Genetic Algoriths, A Bradford Book, The MIT Press, Mass., T. J. Ross, Fuzzy Logic with Engineering Applications, McGraw-Hill Inc., New York, M. Sakawa, K. Kato, H. Sunada, and T. Shibano, Fuzzy prograing for ultiobective - prograing probles through revised genetic algoriths, European Journal of Operational Research, Vol. 97, 997, pp E. Sanchez, T. Shibata, and L. A. Zadeh, ed., Genetic Algoriths and Fuzzy Logic Systes, Soft Coputing Perspectives, World Scientific, C. H. Wang, T. P. Hong, and S. S. Tseng, Integrating fuzzy knowledge by genetic algoriths, IEEE Transactions on Evolutionary Coputation, Vol., 998, pp J. S. Yao and D. C. Lin, Optial fuzzy profit for price in fuzzy sense, Fuzzy Sets

18 58 FENG-TSE LIN AND JING-SHING YAO and Systes, Vol.,, pp J. S. Yao and S. C. Chang, Econoic principle of profit in fuzzy sense, Fuzzy Sets and Systes, Vol.,, pp Feng-Tse Lin ( 林豐澤 ) was born in Taipei, Taiwan. He received the M.S. degree in coputer engineering fro the National Chiao-Tung University, Hsinchu, Taiwan, in 984, and the Ph.D. degree in coputer science and inforation engineering fro the National Taiwan University, Taipei, Taiwan, in 99. Fro 984 to 987, he was a research assistant at Telecounication Laboratories, Chung-Li, Taiwan. Currently, he is an Associate Professor with the Departent of Applied Matheatics, Chinese Culture University, Yanginshan, Taipei, Taiwan. In 997, he was a Visiting Associate Professor, Division of Engineering and Applied Sciences, Harvard University, Cabridge, MA. His current research interests include fuzzy logic and its applications, cobinational optiization probles, genetic algoriths, learning classifier systes, and intelligent agents. Dr. Lin is a eber of Taiwanese Association for Artificial Intelligence (TAAI), Chinese Fuzzy Systes Association Taiwan (CFSAT), Coputer Society of the ROC, and IEEE Systes, Man, and Cybernetics Society. Jing-Shing Yao ( 姚景星 ) was born in Taipei, Taiwan. He received the Ph.D. degree in Matheatics fro National Kyushu University, Japan, in 973. He has been with the Departent of Matheatics, National Taiwan University, Taipei, Taiwan, since 965. He was an associate professor in and as a professor since 967. He headed the Departent of Matheatics, National Taiwan University and also served as Director of Matheatical center, National Science Council, R. O. C., fro 976 to 978. Fro 993 to 998, he was with Departent of Applied Matheatics, Chinese Culture University, Yanginshan, Taipei, Taiwan. He is currently a Professor Eeritus at National Taiwan University, Taipei, Taiwan, and also an Honorary Professor at Manageent College, Takang University, Tasui, Taiwan. His current research interests focus on operation research, fuzzy atheatics and statistics. Dr. Yao has published ore than 5 theoretical papers in various international ournals in the above areas.

LONG-TERM PREDICTIVE VALUE INTERVAL WITH THE FUZZY TIME SERIES

LONG-TERM PREDICTIVE VALUE INTERVAL WITH THE FUZZY TIME SERIES Journal of Marine Science and Technology, Vol 19, No 5, pp 509-513 (2011) 509 LONG-TERM PREDICTIVE VALUE INTERVAL WITH THE FUZZY TIME SERIES Ming-Tao Chou* Key words: fuzzy tie series, fuzzy forecasting,

More information

Finite fields. and we ve used it in various examples and homework problems. In these notes I will introduce more finite fields

Finite fields. and we ve used it in various examples and homework problems. In these notes I will introduce more finite fields Finite fields I talked in class about the field with two eleents F 2 = {, } and we ve used it in various eaples and hoework probles. In these notes I will introduce ore finite fields F p = {,,...,p } for

More information

Soft Computing Techniques Help Assign Weights to Different Factors in Vulnerability Analysis

Soft Computing Techniques Help Assign Weights to Different Factors in Vulnerability Analysis Soft Coputing Techniques Help Assign Weights to Different Factors in Vulnerability Analysis Beverly Rivera 1,2, Irbis Gallegos 1, and Vladik Kreinovich 2 1 Regional Cyber and Energy Security Center RCES

More information

Department of Electronic and Optical Engineering, Ordnance Engineering College, Shijiazhuang, , China

Department of Electronic and Optical Engineering, Ordnance Engineering College, Shijiazhuang, , China 6th International Conference on Machinery, Materials, Environent, Biotechnology and Coputer (MMEBC 06) Solving Multi-Sensor Multi-Target Assignent Proble Based on Copositive Cobat Efficiency and QPSO Algorith

More information

Model Fitting. CURM Background Material, Fall 2014 Dr. Doreen De Leon

Model Fitting. CURM Background Material, Fall 2014 Dr. Doreen De Leon Model Fitting CURM Background Material, Fall 014 Dr. Doreen De Leon 1 Introduction Given a set of data points, we often want to fit a selected odel or type to the data (e.g., we suspect an exponential

More information

Interactive Markov Models of Evolutionary Algorithms

Interactive Markov Models of Evolutionary Algorithms Cleveland State University EngagedScholarship@CSU Electrical Engineering & Coputer Science Faculty Publications Electrical Engineering & Coputer Science Departent 2015 Interactive Markov Models of Evolutionary

More information

e-companion ONLY AVAILABLE IN ELECTRONIC FORM

e-companion ONLY AVAILABLE IN ELECTRONIC FORM OPERATIONS RESEARCH doi 10.1287/opre.1070.0427ec pp. ec1 ec5 e-copanion ONLY AVAILABLE IN ELECTRONIC FORM infors 07 INFORMS Electronic Copanion A Learning Approach for Interactive Marketing to a Custoer

More information

AN IMPROVED FUZZY TIME SERIES THEORY WITH APPLICATIONS IN THE SHANGHAI CONTAINERIZED FREIGHT INDEX

AN IMPROVED FUZZY TIME SERIES THEORY WITH APPLICATIONS IN THE SHANGHAI CONTAINERIZED FREIGHT INDEX Journal of Marine Science and Technology, Vol 5, No, pp 393-398 (01) 393 DOI: 106119/JMST-01-0313-1 AN IMPROVED FUZZY TIME SERIES THEORY WITH APPLICATIONS IN THE SHANGHAI CONTAINERIZED FREIGHT INDEX Ming-Tao

More information

Page 1 Lab 1 Elementary Matrix and Linear Algebra Spring 2011

Page 1 Lab 1 Elementary Matrix and Linear Algebra Spring 2011 Page Lab Eleentary Matri and Linear Algebra Spring 0 Nae Due /03/0 Score /5 Probles through 4 are each worth 4 points.. Go to the Linear Algebra oolkit site ransforing a atri to reduced row echelon for

More information

On the Analysis of the Quantum-inspired Evolutionary Algorithm with a Single Individual

On the Analysis of the Quantum-inspired Evolutionary Algorithm with a Single Individual 6 IEEE Congress on Evolutionary Coputation Sheraton Vancouver Wall Centre Hotel, Vancouver, BC, Canada July 16-1, 6 On the Analysis of the Quantu-inspired Evolutionary Algorith with a Single Individual

More information

Ştefan ŞTEFĂNESCU * is the minimum global value for the function h (x)

Ştefan ŞTEFĂNESCU * is the minimum global value for the function h (x) 7Applying Nelder Mead s Optiization Algorith APPLYING NELDER MEAD S OPTIMIZATION ALGORITHM FOR MULTIPLE GLOBAL MINIMA Abstract Ştefan ŞTEFĂNESCU * The iterative deterinistic optiization ethod could not

More information

Ensemble Based on Data Envelopment Analysis

Ensemble Based on Data Envelopment Analysis Enseble Based on Data Envelopent Analysis So Young Sohn & Hong Choi Departent of Coputer Science & Industrial Systes Engineering, Yonsei University, Seoul, Korea Tel) 82-2-223-404, Fax) 82-2- 364-7807

More information

Non-Parametric Non-Line-of-Sight Identification 1

Non-Parametric Non-Line-of-Sight Identification 1 Non-Paraetric Non-Line-of-Sight Identification Sinan Gezici, Hisashi Kobayashi and H. Vincent Poor Departent of Electrical Engineering School of Engineering and Applied Science Princeton University, Princeton,

More information

MSEC MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL SOLUTION FOR MAINTENANCE AND PERFORMANCE

MSEC MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL SOLUTION FOR MAINTENANCE AND PERFORMANCE Proceeding of the ASME 9 International Manufacturing Science and Engineering Conference MSEC9 October 4-7, 9, West Lafayette, Indiana, USA MSEC9-8466 MODELING OF DEGRADATION PROCESSES TO OBTAIN AN OPTIMAL

More information

Fast Montgomery-like Square Root Computation over GF(2 m ) for All Trinomials

Fast Montgomery-like Square Root Computation over GF(2 m ) for All Trinomials Fast Montgoery-like Square Root Coputation over GF( ) for All Trinoials Yin Li a, Yu Zhang a, a Departent of Coputer Science and Technology, Xinyang Noral University, Henan, P.R.China Abstract This letter

More information

Curious Bounds for Floor Function Sums

Curious Bounds for Floor Function Sums 1 47 6 11 Journal of Integer Sequences, Vol. 1 (018), Article 18.1.8 Curious Bounds for Floor Function Sus Thotsaporn Thanatipanonda and Elaine Wong 1 Science Division Mahidol University International

More information

IAENG International Journal of Computer Science, 42:2, IJCS_42_2_06. Approximation Capabilities of Interpretable Fuzzy Inference Systems

IAENG International Journal of Computer Science, 42:2, IJCS_42_2_06. Approximation Capabilities of Interpretable Fuzzy Inference Systems IAENG International Journal of Coputer Science, 4:, IJCS_4 6 Approxiation Capabilities of Interpretable Fuzzy Inference Systes Hirofui Miyajia, Noritaka Shigei, and Hiroi Miyajia 3 Abstract Many studies

More information

Genetic Algorithm Search for Stent Design Improvements

Genetic Algorithm Search for Stent Design Improvements Genetic Algorith Search for Stent Design Iproveents K. Tesch, M.A. Atherton & M.W. Collins, South Bank University, London, UK Abstract This paper presents an optiisation process for finding iproved stent

More information

A Self-Organizing Model for Logical Regression Jerry Farlow 1 University of Maine. (1900 words)

A Self-Organizing Model for Logical Regression Jerry Farlow 1 University of Maine. (1900 words) 1 A Self-Organizing Model for Logical Regression Jerry Farlow 1 University of Maine (1900 words) Contact: Jerry Farlow Dept of Matheatics Univeristy of Maine Orono, ME 04469 Tel (07) 866-3540 Eail: farlow@ath.uaine.edu

More information

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines Intelligent Systes: Reasoning and Recognition Jaes L. Crowley osig 1 Winter Seester 2018 Lesson 6 27 February 2018 Outline Perceptrons and Support Vector achines Notation...2 Linear odels...3 Lines, Planes

More information

Handwriting Detection Model Based on Four-Dimensional Vector Space Model

Handwriting Detection Model Based on Four-Dimensional Vector Space Model Journal of Matheatics Research; Vol. 10, No. 4; August 2018 ISSN 1916-9795 E-ISSN 1916-9809 Published by Canadian Center of Science and Education Handwriting Detection Model Based on Four-Diensional Vector

More information

Experimental Design For Model Discrimination And Precise Parameter Estimation In WDS Analysis

Experimental Design For Model Discrimination And Precise Parameter Estimation In WDS Analysis City University of New York (CUNY) CUNY Acadeic Works International Conference on Hydroinforatics 8-1-2014 Experiental Design For Model Discriination And Precise Paraeter Estiation In WDS Analysis Giovanna

More information

Uniform Approximation and Bernstein Polynomials with Coefficients in the Unit Interval

Uniform Approximation and Bernstein Polynomials with Coefficients in the Unit Interval Unifor Approxiation and Bernstein Polynoials with Coefficients in the Unit Interval Weiang Qian and Marc D. Riedel Electrical and Coputer Engineering, University of Minnesota 200 Union St. S.E. Minneapolis,

More information

A note on the multiplication of sparse matrices

A note on the multiplication of sparse matrices Cent. Eur. J. Cop. Sci. 41) 2014 1-11 DOI: 10.2478/s13537-014-0201-x Central European Journal of Coputer Science A note on the ultiplication of sparse atrices Research Article Keivan Borna 12, Sohrab Aboozarkhani

More information

TABLE FOR UPPER PERCENTAGE POINTS OF THE LARGEST ROOT OF A DETERMINANTAL EQUATION WITH FIVE ROOTS. By William W. Chen

TABLE FOR UPPER PERCENTAGE POINTS OF THE LARGEST ROOT OF A DETERMINANTAL EQUATION WITH FIVE ROOTS. By William W. Chen TABLE FOR UPPER PERCENTAGE POINTS OF THE LARGEST ROOT OF A DETERMINANTAL EQUATION WITH FIVE ROOTS By Willia W. Chen The distribution of the non-null characteristic roots of a atri derived fro saple observations

More information

Estimation of the Mean of the Exponential Distribution Using Maximum Ranked Set Sampling with Unequal Samples

Estimation of the Mean of the Exponential Distribution Using Maximum Ranked Set Sampling with Unequal Samples Open Journal of Statistics, 4, 4, 64-649 Published Online Septeber 4 in SciRes http//wwwscirporg/ournal/os http//ddoiorg/436/os4486 Estiation of the Mean of the Eponential Distribution Using Maiu Ranked

More information

PAPER Approach to the Unit Maintenance Scheduling Decision Using Risk Assessment and Evolution Programming Techniques

PAPER Approach to the Unit Maintenance Scheduling Decision Using Risk Assessment and Evolution Programming Techniques 1900 PAPER Approach to the Unit Maintenance Scheduling Decision Using Risk Assessent and Evolution Prograing Techniques Chen-Sung CHANG a), Meber SUMMARY This paper applies the Evolutionary Prograing (EP)

More information

FUZZY PARAMETRIC GEOMETRIC PROGRAMMING WITH APPLICATION IN FUZZY EPQ MODEL UNDER FLEXIBILITY AND RELIABILITY CONSIDERATION

FUZZY PARAMETRIC GEOMETRIC PROGRAMMING WITH APPLICATION IN FUZZY EPQ MODEL UNDER FLEXIBILITY AND RELIABILITY CONSIDERATION ISSN 1746-7659, England, UK Journal of Inforation and Coputing Science Vol. 7, No. 3, 1, pp. 3-34 FUZZY PARAMERIC GEOMERIC PROGRAMMING WIH APPLICAION IN FUZZY EPQ MODEL UNDER FLEXIBILIY AND RELIABILIY

More information

C na (1) a=l. c = CO + Clm + CZ TWO-STAGE SAMPLE DESIGN WITH SMALL CLUSTERS. 1. Introduction

C na (1) a=l. c = CO + Clm + CZ TWO-STAGE SAMPLE DESIGN WITH SMALL CLUSTERS. 1. Introduction TWO-STGE SMPLE DESIGN WITH SMLL CLUSTERS Robert G. Clark and David G. Steel School of Matheatics and pplied Statistics, University of Wollongong, NSW 5 ustralia. (robert.clark@abs.gov.au) Key Words: saple

More information

A Division Algorithm Using Bisection Method in Residue Number System

A Division Algorithm Using Bisection Method in Residue Number System International Journal of Coputer, Consuer and Control IJ3C), Vol., No. 03) 59 A Division Algorith Using Bisection Method in Residue Nuber Syste * Chin-Chen Chang and Jen-Ho Yang Abstract. Introduction

More information

Equilibria on the Day-Ahead Electricity Market

Equilibria on the Day-Ahead Electricity Market Equilibria on the Day-Ahead Electricity Market Margarida Carvalho INESC Porto, Portugal Faculdade de Ciências, Universidade do Porto, Portugal argarida.carvalho@dcc.fc.up.pt João Pedro Pedroso INESC Porto,

More information

A Model for the Selection of Internet Service Providers

A Model for the Selection of Internet Service Providers ISSN 0146-4116, Autoatic Control and Coputer Sciences, 2008, Vol. 42, No. 5, pp. 249 254. Allerton Press, Inc., 2008. Original Russian Text I.M. Aliev, 2008, published in Avtoatika i Vychislitel naya Tekhnika,

More information

Support Vector Machine Classification of Uncertain and Imbalanced data using Robust Optimization

Support Vector Machine Classification of Uncertain and Imbalanced data using Robust Optimization Recent Researches in Coputer Science Support Vector Machine Classification of Uncertain and Ibalanced data using Robust Optiization RAGHAV PAT, THEODORE B. TRAFALIS, KASH BARKER School of Industrial Engineering

More information

ASSIGNMENT BOOKLET Bachelor s Degree Programme (B.Sc./B.A./B.Com.) MATHEMATICAL MODELLING

ASSIGNMENT BOOKLET Bachelor s Degree Programme (B.Sc./B.A./B.Com.) MATHEMATICAL MODELLING ASSIGNMENT BOOKLET Bachelor s Degree Prograe (B.Sc./B.A./B.Co.) MTE-14 MATHEMATICAL MODELLING Valid fro 1 st January, 18 to 1 st Deceber, 18 It is copulsory to subit the Assignent before filling in the

More information

On Constant Power Water-filling

On Constant Power Water-filling On Constant Power Water-filling Wei Yu and John M. Cioffi Electrical Engineering Departent Stanford University, Stanford, CA94305, U.S.A. eails: {weiyu,cioffi}@stanford.edu Abstract This paper derives

More information

CHAPTER 8 CONSTRAINED OPTIMIZATION 2: SEQUENTIAL QUADRATIC PROGRAMMING, INTERIOR POINT AND GENERALIZED REDUCED GRADIENT METHODS

CHAPTER 8 CONSTRAINED OPTIMIZATION 2: SEQUENTIAL QUADRATIC PROGRAMMING, INTERIOR POINT AND GENERALIZED REDUCED GRADIENT METHODS CHAPER 8 CONSRAINED OPIMIZAION : SEQUENIAL QUADRAIC PROGRAMMING, INERIOR POIN AND GENERALIZED REDUCED GRADIEN MEHODS 8. Introduction In the previous chapter we eained the necessary and sufficient conditions

More information

A Markov Framework for the Simple Genetic Algorithm

A Markov Framework for the Simple Genetic Algorithm A arkov Fraework for the Siple Genetic Algorith Thoas E. Davis*, Jose C. Principe Electrical Engineering Departent University of Florida, Gainesville, FL 326 *WL/NGS Eglin AFB, FL32542 Abstract This paper

More information

An improved self-adaptive harmony search algorithm for joint replenishment problems

An improved self-adaptive harmony search algorithm for joint replenishment problems An iproved self-adaptive harony search algorith for joint replenishent probles Lin Wang School of Manageent, Huazhong University of Science & Technology zhoulearner@gail.co Xiaojian Zhou School of Manageent,

More information

Keywords: Estimator, Bias, Mean-squared error, normality, generalized Pareto distribution

Keywords: Estimator, Bias, Mean-squared error, normality, generalized Pareto distribution Testing approxiate norality of an estiator using the estiated MSE and bias with an application to the shape paraeter of the generalized Pareto distribution J. Martin van Zyl Abstract In this work the norality

More information

A Simplified Analytical Approach for Efficiency Evaluation of the Weaving Machines with Automatic Filling Repair

A Simplified Analytical Approach for Efficiency Evaluation of the Weaving Machines with Automatic Filling Repair Proceedings of the 6th SEAS International Conference on Siulation, Modelling and Optiization, Lisbon, Portugal, Septeber -4, 006 0 A Siplified Analytical Approach for Efficiency Evaluation of the eaving

More information

Pattern Recognition and Machine Learning. Learning and Evaluation for Pattern Recognition

Pattern Recognition and Machine Learning. Learning and Evaluation for Pattern Recognition Pattern Recognition and Machine Learning Jaes L. Crowley ENSIMAG 3 - MMIS Fall Seester 2017 Lesson 1 4 October 2017 Outline Learning and Evaluation for Pattern Recognition Notation...2 1. The Pattern Recognition

More information

ASSUME a source over an alphabet size m, from which a sequence of n independent samples are drawn. The classical

ASSUME a source over an alphabet size m, from which a sequence of n independent samples are drawn. The classical IEEE TRANSACTIONS ON INFORMATION THEORY Large Alphabet Source Coding using Independent Coponent Analysis Aichai Painsky, Meber, IEEE, Saharon Rosset and Meir Feder, Fellow, IEEE arxiv:67.7v [cs.it] Jul

More information

Pattern Recognition and Machine Learning. Artificial Neural networks

Pattern Recognition and Machine Learning. Artificial Neural networks Pattern Recognition and Machine Learning Jaes L. Crowley ENSIMAG 3 - MMIS Fall Seester 2017 Lessons 7 20 Dec 2017 Outline Artificial Neural networks Notation...2 Introduction...3 Key Equations... 3 Artificial

More information

Figure 1: Equivalent electric (RC) circuit of a neurons membrane

Figure 1: Equivalent electric (RC) circuit of a neurons membrane Exercise: Leaky integrate and fire odel of neural spike generation This exercise investigates a siplified odel of how neurons spike in response to current inputs, one of the ost fundaental properties of

More information

Stochastic Subgradient Methods

Stochastic Subgradient Methods Stochastic Subgradient Methods Lingjie Weng Yutian Chen Bren School of Inforation and Coputer Science University of California, Irvine {wengl, yutianc}@ics.uci.edu Abstract Stochastic subgradient ethods

More information

When Short Runs Beat Long Runs

When Short Runs Beat Long Runs When Short Runs Beat Long Runs Sean Luke George Mason University http://www.cs.gu.edu/ sean/ Abstract What will yield the best results: doing one run n generations long or doing runs n/ generations long

More information

A general forulation of the cross-nested logit odel Michel Bierlaire, Dpt of Matheatics, EPFL, Lausanne Phone: Fax:

A general forulation of the cross-nested logit odel Michel Bierlaire, Dpt of Matheatics, EPFL, Lausanne Phone: Fax: A general forulation of the cross-nested logit odel Michel Bierlaire, EPFL Conference paper STRC 2001 Session: Choices A general forulation of the cross-nested logit odel Michel Bierlaire, Dpt of Matheatics,

More information

Block designs and statistics

Block designs and statistics Bloc designs and statistics Notes for Math 447 May 3, 2011 The ain paraeters of a bloc design are nuber of varieties v, bloc size, nuber of blocs b. A design is built on a set of v eleents. Each eleent

More information

Identical Maximum Likelihood State Estimation Based on Incremental Finite Mixture Model in PHD Filter

Identical Maximum Likelihood State Estimation Based on Incremental Finite Mixture Model in PHD Filter Identical Maxiu Lielihood State Estiation Based on Increental Finite Mixture Model in PHD Filter Gang Wu Eail: xjtuwugang@gail.co Jing Liu Eail: elelj20080730@ail.xjtu.edu.cn Chongzhao Han Eail: czhan@ail.xjtu.edu.cn

More information

THE KALMAN FILTER: A LOOK BEHIND THE SCENE

THE KALMAN FILTER: A LOOK BEHIND THE SCENE HE KALMA FILER: A LOOK BEHID HE SCEE R.E. Deain School of Matheatical and Geospatial Sciences, RMI University eail: rod.deain@rit.edu.au Presented at the Victorian Regional Survey Conference, Mildura,

More information

On Rough Interval Three Level Large Scale Quadratic Integer Programming Problem

On Rough Interval Three Level Large Scale Quadratic Integer Programming Problem J. Stat. Appl. Pro. 6, No. 2, 305-318 2017) 305 Journal of Statistics Applications & Probability An International Journal http://dx.doi.org/10.18576/jsap/060206 On Rough Interval Three evel arge Scale

More information

World's largest Science, Technology & Medicine Open Access book publisher

World's largest Science, Technology & Medicine Open Access book publisher PUBLISHED BY World's largest Science, Technology & Medicine Open Access book publisher 2750+ OPEN ACCESS BOOKS 95,000+ INTERNATIONAL AUTHORS AND EDITORS 88+ MILLION DOWNLOADS BOOKS DELIVERED TO 5 COUNTRIES

More information

High-Speed Smooth Cyclical Motion of a Hydraulic Cylinder

High-Speed Smooth Cyclical Motion of a Hydraulic Cylinder 846 1 th International Conference on Fleible Autoation and Intelligent Manufacturing 00, Dresden, Gerany High-Speed Sooth Cyclical Motion of a Hydraulic Cylinder Nebojsa I. Jaksic Departent of Industrial

More information

AN APPLICATION OF CUBIC B-SPLINE FINITE ELEMENT METHOD FOR THE BURGERS EQUATION

AN APPLICATION OF CUBIC B-SPLINE FINITE ELEMENT METHOD FOR THE BURGERS EQUATION Aksan, E..: An Applıcatıon of Cubıc B-Splıne Fınıte Eleent Method for... THERMAL SCIECE: Year 8, Vol., Suppl., pp. S95-S S95 A APPLICATIO OF CBIC B-SPLIE FIITE ELEMET METHOD FOR THE BRGERS EQATIO by Eine

More information

Convolutional Codes. Lecture Notes 8: Trellis Codes. Example: K=3,M=2, rate 1/2 code. Figure 95: Convolutional Encoder

Convolutional Codes. Lecture Notes 8: Trellis Codes. Example: K=3,M=2, rate 1/2 code. Figure 95: Convolutional Encoder Convolutional Codes Lecture Notes 8: Trellis Codes In this lecture we discuss construction of signals via a trellis. That is, signals are constructed by labeling the branches of an infinite trellis with

More information

COS 424: Interacting with Data. Written Exercises

COS 424: Interacting with Data. Written Exercises COS 424: Interacting with Data Hoework #4 Spring 2007 Regression Due: Wednesday, April 18 Written Exercises See the course website for iportant inforation about collaboration and late policies, as well

More information

Using EM To Estimate A Probablity Density With A Mixture Of Gaussians

Using EM To Estimate A Probablity Density With A Mixture Of Gaussians Using EM To Estiate A Probablity Density With A Mixture Of Gaussians Aaron A. D Souza adsouza@usc.edu Introduction The proble we are trying to address in this note is siple. Given a set of data points

More information

MANY physical structures can conveniently be modelled

MANY physical structures can conveniently be modelled Proceedings of the World Congress on Engineering Coputer Science 2017 Vol II Roly r-orthogonal (g, f)-factorizations in Networks Sizhong Zhou Abstract Let G (V (G), E(G)) be a graph, where V (G) E(G) denote

More information

Combining Classifiers

Combining Classifiers Cobining Classifiers Generic ethods of generating and cobining ultiple classifiers Bagging Boosting References: Duda, Hart & Stork, pg 475-480. Hastie, Tibsharini, Friedan, pg 246-256 and Chapter 10. http://www.boosting.org/

More information

Inverted Pendulum control with pole assignment, LQR and multiple layers sliding mode control

Inverted Pendulum control with pole assignment, LQR and multiple layers sliding mode control J. Basic. Appl. Sci. Res., 3(1s)363-368, 013 013, TetRoad Publication ISSN 090-4304 Journal of Basic and Applied Scientific Research www.tetroad.co Inverted Pendulu control with pole assignent, LQR and

More information

Genetic Quantum Algorithm and its Application to Combinatorial Optimization Problem

Genetic Quantum Algorithm and its Application to Combinatorial Optimization Problem Genetic Quantu Algorith and its Application to Cobinatorial Optiization Proble Kuk-Hyun Han Dept. of Electrical Engineering, KAIST, 373-, Kusong-dong Yusong-gu Taejon, 305-70, Republic of Korea khhan@vivaldi.kaist.ac.kr

More information

Multi-Dimensional Hegselmann-Krause Dynamics

Multi-Dimensional Hegselmann-Krause Dynamics Multi-Diensional Hegselann-Krause Dynaics A. Nedić Industrial and Enterprise Systes Engineering Dept. University of Illinois Urbana, IL 680 angelia@illinois.edu B. Touri Coordinated Science Laboratory

More information

A New Approach to Solving Dynamic Traveling Salesman Problems

A New Approach to Solving Dynamic Traveling Salesman Problems A New Approach to Solving Dynaic Traveling Salesan Probles Changhe Li 1 Ming Yang 1 Lishan Kang 1 1 China University of Geosciences(Wuhan) School of Coputer 4374 Wuhan,P.R.China lchwfx@yahoo.co,yanging72@gail.co,ang_whu@yahoo.co

More information

arxiv: v1 [cs.ds] 3 Feb 2014

arxiv: v1 [cs.ds] 3 Feb 2014 arxiv:40.043v [cs.ds] 3 Feb 04 A Bound on the Expected Optiality of Rando Feasible Solutions to Cobinatorial Optiization Probles Evan A. Sultani The Johns Hopins University APL evan@sultani.co http://www.sultani.co/

More information

Sharp Time Data Tradeoffs for Linear Inverse Problems

Sharp Time Data Tradeoffs for Linear Inverse Problems Sharp Tie Data Tradeoffs for Linear Inverse Probles Saet Oyak Benjain Recht Mahdi Soltanolkotabi January 016 Abstract In this paper we characterize sharp tie-data tradeoffs for optiization probles used

More information

Constrained Consensus and Optimization in Multi-Agent Networks arxiv: v2 [math.oc] 17 Dec 2008

Constrained Consensus and Optimization in Multi-Agent Networks arxiv: v2 [math.oc] 17 Dec 2008 LIDS Report 2779 1 Constrained Consensus and Optiization in Multi-Agent Networks arxiv:0802.3922v2 [ath.oc] 17 Dec 2008 Angelia Nedić, Asuan Ozdaglar, and Pablo A. Parrilo February 15, 2013 Abstract We

More information

Warning System of Dangerous Chemical Gas in Factory Based on Wireless Sensor Network

Warning System of Dangerous Chemical Gas in Factory Based on Wireless Sensor Network 565 A publication of CHEMICAL ENGINEERING TRANSACTIONS VOL. 59, 07 Guest Editors: Zhuo Yang, Junie Ba, Jing Pan Copyright 07, AIDIC Servizi S.r.l. ISBN 978-88-95608-49-5; ISSN 83-96 The Italian Association

More information

Solving initial value problems by residual power series method

Solving initial value problems by residual power series method Theoretical Matheatics & Applications, vol.3, no.1, 13, 199-1 ISSN: 179-9687 (print), 179-979 (online) Scienpress Ltd, 13 Solving initial value probles by residual power series ethod Mohaed H. Al-Sadi

More information

On Poset Merging. 1 Introduction. Peter Chen Guoli Ding Steve Seiden. Keywords: Merging, Partial Order, Lower Bounds. AMS Classification: 68W40

On Poset Merging. 1 Introduction. Peter Chen Guoli Ding Steve Seiden. Keywords: Merging, Partial Order, Lower Bounds. AMS Classification: 68W40 On Poset Merging Peter Chen Guoli Ding Steve Seiden Abstract We consider the follow poset erging proble: Let X and Y be two subsets of a partially ordered set S. Given coplete inforation about the ordering

More information

Construction of an index by maximization of the sum of its absolute correlation coefficients with the constituent variables

Construction of an index by maximization of the sum of its absolute correlation coefficients with the constituent variables Construction of an index by axiization of the su of its absolute correlation coefficients with the constituent variables SK Mishra Departent of Econoics North-Eastern Hill University Shillong (India) I.

More information

A Better Algorithm For an Ancient Scheduling Problem. David R. Karger Steven J. Phillips Eric Torng. Department of Computer Science

A Better Algorithm For an Ancient Scheduling Problem. David R. Karger Steven J. Phillips Eric Torng. Department of Computer Science A Better Algorith For an Ancient Scheduling Proble David R. Karger Steven J. Phillips Eric Torng Departent of Coputer Science Stanford University Stanford, CA 9435-4 Abstract One of the oldest and siplest

More information

Derivative at a point

Derivative at a point Roberto s Notes on Differential Calculus Capter : Definition of derivative Section Derivative at a point Wat you need to know already: Te concept of liit and basic etods for coputing liits. Wat you can

More information

ON REGULARITY, TRANSITIVITY, AND ERGODIC PRINCIPLE FOR QUADRATIC STOCHASTIC VOLTERRA OPERATORS MANSOOR SABUROV

ON REGULARITY, TRANSITIVITY, AND ERGODIC PRINCIPLE FOR QUADRATIC STOCHASTIC VOLTERRA OPERATORS MANSOOR SABUROV ON REGULARITY TRANSITIVITY AND ERGODIC PRINCIPLE FOR QUADRATIC STOCHASTIC VOLTERRA OPERATORS MANSOOR SABUROV Departent of Coputational & Theoretical Sciences Faculty of Science International Islaic University

More information

A NEW ROBUST AND EFFICIENT ESTIMATOR FOR ILL-CONDITIONED LINEAR INVERSE PROBLEMS WITH OUTLIERS

A NEW ROBUST AND EFFICIENT ESTIMATOR FOR ILL-CONDITIONED LINEAR INVERSE PROBLEMS WITH OUTLIERS A NEW ROBUST AND EFFICIENT ESTIMATOR FOR ILL-CONDITIONED LINEAR INVERSE PROBLEMS WITH OUTLIERS Marta Martinez-Caara 1, Michael Mua 2, Abdelhak M. Zoubir 2, Martin Vetterli 1 1 School of Coputer and Counication

More information

INTEGRATIVE COOPERATIVE APPROACH FOR SOLVING PERMUTATION FLOWSHOP SCHEDULING PROBLEM WITH SEQUENCE DEPENDENT FAMILY SETUP TIMES

INTEGRATIVE COOPERATIVE APPROACH FOR SOLVING PERMUTATION FLOWSHOP SCHEDULING PROBLEM WITH SEQUENCE DEPENDENT FAMILY SETUP TIMES 8 th International Conference of Modeling and Siulation - MOSIM 10 - May 10-12, 2010 - Haaet - Tunisia Evaluation and optiization of innovative production systes of goods and services INTEGRATIVE COOPERATIVE

More information

An Approximate Model for the Theoretical Prediction of the Velocity Increase in the Intermediate Ballistics Period

An Approximate Model for the Theoretical Prediction of the Velocity Increase in the Intermediate Ballistics Period An Approxiate Model for the Theoretical Prediction of the Velocity... 77 Central European Journal of Energetic Materials, 205, 2(), 77-88 ISSN 2353-843 An Approxiate Model for the Theoretical Prediction

More information

Solving Fuzzy Linear Fractional Programming. Problem Using Metric Distance Ranking

Solving Fuzzy Linear Fractional Programming. Problem Using Metric Distance Ranking pplied Matheatical Sciences, Vol. 6, 0, no. 6, 75-85 Solving Fuzzy inear Fractional Prograing Proble Using Metric Distance anking l. Nachaai Kongu Engineering College, Erode, Tailnadu, India nachusenthilnathan@gail.co

More information

IN modern society that various systems have become more

IN modern society that various systems have become more Developent of Reliability Function in -Coponent Standby Redundant Syste with Priority Based on Maxiu Entropy Principle Ryosuke Hirata, Ikuo Arizono, Ryosuke Toohiro, Satoshi Oigawa, and Yasuhiko Takeoto

More information

Bootstrapping Dependent Data

Bootstrapping Dependent Data Bootstrapping Dependent Data One of the key issues confronting bootstrap resapling approxiations is how to deal with dependent data. Consider a sequence fx t g n t= of dependent rando variables. Clearly

More information

MULTIAGENT Resource Allocation (MARA) is the

MULTIAGENT Resource Allocation (MARA) is the EDIC RESEARCH PROPOSAL 1 Designing Negotiation Protocols for Utility Maxiization in Multiagent Resource Allocation Tri Kurniawan Wijaya LSIR, I&C, EPFL Abstract Resource allocation is one of the ain concerns

More information

Nonlinear Stabilization of a Spherical Particle Trapped in an Optical Tweezer

Nonlinear Stabilization of a Spherical Particle Trapped in an Optical Tweezer Nonlinear Stabilization of a Spherical Particle Trapped in an Optical Tweezer Aruna Ranaweera ranawera@engineering.ucsb.edu Bassa Baieh baieh@engineering.ucsb.edu Andrew R. Teel teel@ece.ucsb.edu Departent

More information

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation Course Notes for EE227C (Spring 2018): Convex Optiization and Approxiation Instructor: Moritz Hardt Eail: hardt+ee227c@berkeley.edu Graduate Instructor: Max Sichowitz Eail: sichow+ee227c@berkeley.edu October

More information

EXPLICIT CONGRUENCES FOR EULER POLYNOMIALS

EXPLICIT CONGRUENCES FOR EULER POLYNOMIALS EXPLICIT CONGRUENCES FOR EULER POLYNOMIALS Zhi-Wei Sun Departent of Matheatics, Nanjing University Nanjing 10093, People s Republic of China zwsun@nju.edu.cn Abstract In this paper we establish soe explicit

More information

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and This article appeared in a ournal published by Elsevier. The attached copy is furnished to the author for internal non-coercial research and education use, including for instruction at the authors institution

More information

Numerical Studies of a Nonlinear Heat Equation with Square Root Reaction Term

Numerical Studies of a Nonlinear Heat Equation with Square Root Reaction Term Nuerical Studies of a Nonlinear Heat Equation with Square Root Reaction Ter Ron Bucire, 1 Karl McMurtry, 1 Ronald E. Micens 2 1 Matheatics Departent, Occidental College, Los Angeles, California 90041 2

More information

Pattern Recognition and Machine Learning. Artificial Neural networks

Pattern Recognition and Machine Learning. Artificial Neural networks Pattern Recognition and Machine Learning Jaes L. Crowley ENSIMAG 3 - MMIS Fall Seester 2016 Lessons 7 14 Dec 2016 Outline Artificial Neural networks Notation...2 1. Introduction...3... 3 The Artificial

More information

Intelligent Systems: Reasoning and Recognition. Artificial Neural Networks

Intelligent Systems: Reasoning and Recognition. Artificial Neural Networks Intelligent Systes: Reasoning and Recognition Jaes L. Crowley MOSIG M1 Winter Seester 2018 Lesson 7 1 March 2018 Outline Artificial Neural Networks Notation...2 Introduction...3 Key Equations... 3 Artificial

More information

Symbolic Analysis as Universal Tool for Deriving Properties of Non-linear Algorithms Case study of EM Algorithm

Symbolic Analysis as Universal Tool for Deriving Properties of Non-linear Algorithms Case study of EM Algorithm Acta Polytechnica Hungarica Vol., No., 04 Sybolic Analysis as Universal Tool for Deriving Properties of Non-linear Algoriths Case study of EM Algorith Vladiir Mladenović, Miroslav Lutovac, Dana Porrat

More information

EMPIRICAL COMPLEXITY ANALYSIS OF A MILP-APPROACH FOR OPTIMIZATION OF HYBRID SYSTEMS

EMPIRICAL COMPLEXITY ANALYSIS OF A MILP-APPROACH FOR OPTIMIZATION OF HYBRID SYSTEMS EMPIRICAL COMPLEXITY ANALYSIS OF A MILP-APPROACH FOR OPTIMIZATION OF HYBRID SYSTEMS Jochen Till, Sebastian Engell, Sebastian Panek, and Olaf Stursberg Process Control Lab (CT-AST), University of Dortund,

More information

a a a a a a a m a b a b

a a a a a a a m a b a b Algebra / Trig Final Exa Study Guide (Fall Seester) Moncada/Dunphy Inforation About the Final Exa The final exa is cuulative, covering Appendix A (A.1-A.5) and Chapter 1. All probles will be ultiple choice

More information

2 Q 10. Likewise, in case of multiple particles, the corresponding density in 2 must be averaged over all

2 Q 10. Likewise, in case of multiple particles, the corresponding density in 2 must be averaged over all Lecture 6 Introduction to kinetic theory of plasa waves Introduction to kinetic theory So far we have been odeling plasa dynaics using fluid equations. The assuption has been that the pressure can be either

More information

Applications of Maximum Entropy Principle in Formulating Solution of Constrained Non-Linear Programming Problem

Applications of Maximum Entropy Principle in Formulating Solution of Constrained Non-Linear Programming Problem Journal of Statistical and Econoetric Methods, vol.4, no., 05, 45-60 ISSN: 4-084 (print), 4-076 (online) Scienpress Ltd, 05 Applications of Maiu Entropy Principle in Forulatin Solution of Constrained Non-Linear

More information

Homotopy Analysis Method for Solving Fuzzy Integro-Differential Equations

Homotopy Analysis Method for Solving Fuzzy Integro-Differential Equations Modern Applied Science; Vol. 7 No. 3; 23 ISSN 93-844 E-ISSN 93-82 Published by Canadian Center of Science Education Hootopy Analysis Method for Solving Fuzzy Integro-Differential Equations Ean A. Hussain

More information

Lost-Sales Problems with Stochastic Lead Times: Convexity Results for Base-Stock Policies

Lost-Sales Problems with Stochastic Lead Times: Convexity Results for Base-Stock Policies OPERATIONS RESEARCH Vol. 52, No. 5, Septeber October 2004, pp. 795 803 issn 0030-364X eissn 1526-5463 04 5205 0795 infors doi 10.1287/opre.1040.0130 2004 INFORMS TECHNICAL NOTE Lost-Sales Probles with

More information

Linguistic majorities with difference in support

Linguistic majorities with difference in support Linguistic ajorities with difference in support Patrizia Pérez-Asurendi a, Francisco Chiclana b,c, a PRESAD Research Group, SEED Research Group, IMUVA, Universidad de Valladolid, Valladolid, Spain b Centre

More information

Outperforming the Competition in Multi-Unit Sealed Bid Auctions

Outperforming the Competition in Multi-Unit Sealed Bid Auctions Outperforing the Copetition in Multi-Unit Sealed Bid Auctions ABSTRACT Ioannis A. Vetsikas School of Electronics and Coputer Science University of Southapton Southapton SO17 1BJ, UK iv@ecs.soton.ac.uk

More information

1 Bounding the Margin

1 Bounding the Margin COS 511: Theoretical Machine Learning Lecturer: Rob Schapire Lecture #12 Scribe: Jian Min Si March 14, 2013 1 Bounding the Margin We are continuing the proof of a bound on the generalization error of AdaBoost

More information

Algorithms for parallel processor scheduling with distinct due windows and unit-time jobs

Algorithms for parallel processor scheduling with distinct due windows and unit-time jobs BULLETIN OF THE POLISH ACADEMY OF SCIENCES TECHNICAL SCIENCES Vol. 57, No. 3, 2009 Algoriths for parallel processor scheduling with distinct due windows and unit-tie obs A. JANIAK 1, W.A. JANIAK 2, and

More information

A Note on Scheduling Tall/Small Multiprocessor Tasks with Unit Processing Time to Minimize Maximum Tardiness

A Note on Scheduling Tall/Small Multiprocessor Tasks with Unit Processing Time to Minimize Maximum Tardiness A Note on Scheduling Tall/Sall Multiprocessor Tasks with Unit Processing Tie to Miniize Maxiu Tardiness Philippe Baptiste and Baruch Schieber IBM T.J. Watson Research Center P.O. Box 218, Yorktown Heights,

More information

Estimating Parameters for a Gaussian pdf

Estimating Parameters for a Gaussian pdf Pattern Recognition and achine Learning Jaes L. Crowley ENSIAG 3 IS First Seester 00/0 Lesson 5 7 Noveber 00 Contents Estiating Paraeters for a Gaussian pdf Notation... The Pattern Recognition Proble...3

More information