Parallelized Side-Channel Attack Resisted Scalar Multiplication Using q-based Addition-Subtraction k-chains

Size: px
Start display at page:

Download "Parallelized Side-Channel Attack Resisted Scalar Multiplication Using q-based Addition-Subtraction k-chains"

Transcription

1 Parallelized Side-Channel Atta Resisted Salar Multipliation Using -Based Addition-Subtration -hains Kittiphop Phalaarn Department of Computer Engineering Chulalongorn University Kittiphon Phalaarn Department of Computer Engineering Chulalongorn University Vorapong Suppaitpaisarn Department of Computer Siene The University of Toyo Abstrat This paper presents parallel salar multipliation tehniues for ellipti urve ryptography using -based additionsubtration -hain whih an also effetively resist side-hannel atta. Many tehniues have been disussed to improve salar multipliation, for example, double-and-add, NAF, w-naf, addition hain and addition-subtration hain. However, these tehniues annot resist side-hannel atta. Montgomery ladder, random w-naf and uniform operation tehniues are also widely used to prevent side-hannel atta, but their operations are not effiient enough omparing to those with no side-hannel atta prevention. We have found a new way to use -hain for this purpose. In this paper, we extend the definition of -hain to -based addition-subtration -hain and modify an algorithm proposed by Jarvinen et al. to generate the -based additionsubtration -hain. We show the upper and lower bounds of its length whih lead to the omputation time using the new hain tehniues. The hain tehniues are used to redue the ost of salar multipliation in parallel ways. Comparing to w- NAF, whih is faster than double-and-add and Montgomery ladder tehniue, the maximum omputation time of our - based addition-subtration -hain tehniues an have up to 5.9% less addition osts using only 3 parallel omputing ores. We also disuss on the optimization for multiple operand point addition using hybrid-double multiplier whih is proposed by Azarderahsh and Reyhani-Masoleh. The proposed parallel hain tehniues an also tolerate side-hannel atta effiiently. eywords Information and Communiation Seurity; Effiient Implementations; Parallel Algorithms; Ellipti Curve Cryptography; Salar Multipliation; -Chain; Side-Channel Atta Countermeasure I. INTRODUCTION Today s ryptosystem reuires faster omputation for more seurity. The ost of salar multipliation is very important for ellipti urve ryptography. Many wors try to derease this alulation ost, for example, double-and-add tehniue f. [] whih uses simple binary representation. It an be alulated faster using NAF [] and w-naf [3] tehniue. NAF tehniue uses binary representation with digit set {, 0, } and w-naf representation uses larger digit set. Another tehniue used today is wr-naf [4] whih generalizes w-naf using base r representation. There is another ind of tehniues whih is different from double-and-add. Addition hain f. [5] and additionsubtration hain a/s hain [6] tehniues are used for the This paper has been aepted for publiation at proeedings of the Fourth International Symposium on Computing and Networing CANDAR 06, whih is published by IEEE. It has been further edited by IEEE, and the final version is appearing at same purpose as the general ases for double-and-add and NAF [7], [8]. There are many variations on the addition hain and a/s hain suh as addition-multipliation hain [9], - addition hain [0] and -hain []. Currently, -hain is only used to improve finite field element inversion [] and we are interested in using this -hain to improve salar multipliation. One important problem for salar multipliation is sidehannel atta. The power used in eah alulation step an be measured and give information on whih type of operation is being alulated. This is alled power atta []. Time of eah omputation step an also be measured for the same purpose. The latter atta is alled timing atta [3]. The seuenes of the monitored operations an tell whih salar has been used to multiply. The multipliation tehniues mentioned earlier are not tolerate side-hannel atta. Some studies try to add some random dummy operations to hide the real power and time used, but these operations inrease the alulation osts whih are now lose to the worst ase of eah tehniue. Montgomery ladder [4], random w-naf [5] and uniform operation tehniues [6] are widely used to prevent side-hannel atta, but their operations are not effiient enough omparing to those with no side-hannel atta prevention. A. Contribution of this paper To effiiently resist the side-hannel atta, we have found a new way to use -hain for salar multipliation. In this paper, we generalize the definition of -hain to addition-subtration -hain a/s -hain and -based addition-subtration -hain -based a/s -hain. We modify the algorithm from [] to generate the -based a/s -hain. From the algorithms, we found that summation of all digits in the number representation affets the length of the hain. This summation is alled relaxed ost [7]. We use the relaxed redued representation s properties from [7] to prove the theorem of the maximum relaxed ost of the representation. This leads to the upper and lower bounds of the hain s length and the worst ase of the omputation time for salar multipliation using -based a/s -hain tehniue. We propose a parallel tehniue of -based a/s 4-hain and generalize to -based a/s -hain for any positive integers and greater than. We also propose an optimization tehniue for multiple operand point addition for twisted

2 Edwards urves [8] by using hybrid-double multiplier [9], [0]. We ompare eah tehniue using worst ase omputation time due to the properties of side-hannel atta. Comparing to the best w-naf with w = 5 [4], whih is faster than doubleand-add and Montgomery ladder tehniue, the maximum omputation time of our -based addition-subtration -hain tehniue using 8-bit salar is slightly smaller with omputing ores. For large salar, our hain tehniue an have up to 5.9% less additions using only 3 parallel omputing ores, but if using omputing ores, our tehniue an have up to.% more additions. Finally, we disuss on the appropriate value of in -based a/s -hain and have some seurity analysis of the proposed hain tehniues on side-hannel atta. B. Paper Organization The paper is strutured as follows: In Setion II, we present some previous wors on this topi, inluding relaxed redued representation, -hain, double-and-add tehniue, w-naf tehniue and how to alulate the ost of salar multipliation. In Setion III, we give definitions on a/s -hain and -based a/s -hain. We also modify the -hain algorithm to generate -based a/s -hain. In Setion IV, we prove the maximum relaxed ost of signed digit representations and propose a theorem on the upper and lower bounds of -based a/s -hain length. This theorem leads us to the omputation time using the hain tehniues. In Setion V, we propose some parallel tehniues and give an optimization on multiple operand point addition using hybrid-double multiplier [9], [0]. We ompare the osts of our tehniues in Setion VI and ompare to other well-nown tehniues, w-naf and Montgomery ladder. In the same setion, we also disuss on the value of and give some seurity analysis of the proposed hain tehniues on side-hannel atta. Finally, we onlude our paper in Setion VII. II. PRELIMINARIES In this setion, we present some previous wors, inluding relaxed redued representation [7], -hain [], double-andadd tehniue f. [] and w-naf tehniue [3] for salar multipliation. A. Relaxed Redued Representation We use the definitions and theorems from [7], and we will fous on relaxed ost beause they are more ompatible with our appliations. Definition set of all representations [7]. The set of all representations of an integer n in base is defined as { R n := ε = ε 0,..., ε l l N, ε i Z, ε l 0, n = } l ε i i and for eah ε R n, l is alled the length of the representation ε and written as ε. Example. Given n = 5 and = 4, we have some representations 3, 3 4,, 0, 4,, 4 R 4 5 with length, and, respetively. Definition 3 relaxed ost of the representation [7]. The relaxed ost of representation ε R n is defined as ε = ε 0,..., ε l := l ε i. Definition 4 relaxed redued representation. The representation ε R n is relaxed redued if ε arg min ε. ε R n Example 5. From Example, the relaxed osts of the example representations are 6, and, respetively. It is easy to see that, 0, 4 is relaxed redued representation of R 4 5. Definition 6 redution rules [7]. Given ε R n, f : R n R n is defined as follows. The rules are applied in this order at the first ourrene from the left side until ε first hanges and return the result. If there is no math, fε = ε. For η 0 >, η η Z, and u := 0 : η 0, η η 0 u signη 0, η + u signη 0. For η 0 < 0 and :, η 0, η 0 +. For η 0 > 0 and :, η 0, η 0. For η 0 < and :,, η 0 For < η 0 and :,, η 0, +, η 0 +.,, η 0. Theorem 7 [7]. The redution rules an be applied to the representation ε R n many times until there is no hange. The representation ε is relaxed redued if and only if fε = ε. Theorem 8 Average Relaxed Cost of Relaxed Redued Representation [7]. Given positive integer m and base, the average relaxed ost of the relaxed redued representations of the integer 0,..., m is m ε R i m = m + O if, m + O if. Algorithm has been proposed in [7] to generate relaxed redued representation of an integer.

3 Algorithm [7] generating base relaxed redued representation of positive integer n Input: n > 0, Output: relaxed redued representation ε R n : ε : m n 3: while m > 0 do 4: a m mod 5: if a > / or a = / and m mod / then 6: a a 7: end if 8: m m a/ 9: onatenate a to the right end of ε 0: end while : return ε To generate redued representation the relaxed redued representation with the shortest length, hange the ondition in the if statement in line 5 to not a < / or a = / and m mod < / or m = + / or m = / + /. B. -Chain We give the definition of -hain from []. Previous wors on addition hain, a/s hain and -hain an be found in [5], [6], []. Definition 9 -hain []. Given positive integer and n, a -hain for n is defined as v = [v0,..., vs] with v0 =, vs = n, and vi = h=0 vi h for all 0 < i s and i h < i with v = 0. The length of the hain is eual to s. The algorithm for generating -hain from traditional base representation is given in [] and we show it below. The same paper also proved the upper and lower bounds of the length of the optimal -hain by using the given algorithm. We state it in Theorem 0. Remar: At line 3 of Algorithm, the length of the output -hain will be the same regardless to the order of x i being hosen. Theorem 0 bound of -hain s length []. Given positive integers and n, The upper and lower bounds of the optimal length s of -hain for n are ε log n s log n + where ε R n is a traditional base representation, i.e., using digit set {0,,..., }. C. Current Tehniues on Salar Multipliation To alulate np from the given integer n and ellipti urve point P, the most simple tehniue used is double-and-add f. []. Assume that we have the traditional binary representation Algorithm [] Algorithm for generating -hain for integer n Input:, n = l n i i = n 0,..., n l traditional representation Output: -hain for n : vi i for 0 i l : j 0 3: for all x i 0 in x = {n 0, n,..., n l, n l } do 4: t j, t j+,..., t j+ xi i,...i 5: j j + x i 6: end for 7: t j, t j+,..., t j/,,..., 8: for i = 0 to j/ do 9: vl + i + vl + i + vt i+ + vt i++ 0: + + vt i++ : end for : return v of n, n 0,..., n l, we an ompute np from n l P + n l P + + n 0 P Example. The binary representation of 0 is 0,, 0,. Thus, we an alulate 0P from P +0P +P +0P. To see the alulation ost of this tehniue, we define the ost of point doubling as D and point addition as A. Given positive integer n and traditional binary representation ε R n, the ost for salar multipliation with double-and-add tehniue is defined as Cost n,double = D ε + A ε. Remar: For twisted Edwards urves [8], whih is the urrent fastest urve, the ost for point addition A is now reorded at M where M is the ost for field elements multipliation. The ost for point doubling D is reorded at 7M. Given n as an m-bit integer where m := log n +, We define average and maximum osts for salar multipliation as follows. Avg m,tehniue = m Cost i,tehniue m Max m,tehniue = max 0 i m Cost i,tehniue Given a positive integer n with m bits, the length of its binary representation is exatly m. From Subsetion II-A, we have average relaxed ost euals to m + O and it is easy to see that the maximum relaxed ost euals to m + O. Thus, the final average and maximum osts for double-andadd tehniue are Avg m,double = D m + A m + O, Max m,double = D m + A m + O. Another tehniue widely used is w-naf tehniue. The w-naf tehniue was introdued in [3] and generalized to wr-naf in [4]. The paper stated that given the window size

4 r wr + w and base r, the nonzero density of the base r representation is and the number of points that have to be preomputed is rw r w. It is easy to see that the worst ase for nonzero density is w. With wr-naf, we use number of nonzero digits instead of relaxed ost. To alulate np, we define m := log n +. The average and maximum osts for wr-naf tehniue are similar to those for double-and-add and an be written as follows. Avg m,wr-naf = R m log r r m +A wr + log r + rw r w Max m,wr-naf = R m log r + A m w log r + rw r w, when R is the alulation ost for multiplying the point by r. III. -BASED ADDITION-SUBTRACTION -CHAIN We generalize the definition of -hain to a/s -hain and we will extend the definition to -based a/s -hain. Definition addition-subtration -hain. Given positive integer and n, an addition-subtration -hain a/s -hain for n is defined as v = [v0,..., vs] with v0 =, vs = n, and vi = h=0 ±vi h for all 0 < i s and i h < i with v = 0. Example 3. An a/s 3-hain for 8 is [, 3, 9, 8] 3 beause 3 = v+v+v = ++, 9 = v+v+v = and 8 = v3 v + v = Other examples are [, 3, 7, 8] 3 and [,, 6, 0, 8] 3. Definition 4 -based addition-subtration -hain. An addition-subtration -hain v is a -based additionsubtration -hain -based a/s -hain, if and only if, and there exists an integer j suh that for all 0 < i j, vi = i. The [v0,..., vj] part is alled power part and [vj +,..., vs] part is alled addition part. Example 5. A -based a/s 3-hain for 0 is [,, 4, 8, 0] 3. The power part is [,, 4, 8] 3 = [ 0,,, 3 ] 3 and the addition part is [0] 3 where 0 = v3 + v + v = Other examples are [,, 4, 0] 3, [,, 3, 9, 0] 3 and [,, 4, 8, 6, 0] 3. You should notie that [, 3, 9, 0] 3 is an a/s 3-hain but not a -based a/s 3-hain. Next, we modify Algorithm to Algorithm 3 for generating -based a/s -hain. If the desired output is just a/s -hain, you an use =., Algorithm 3 Modified algorithm for generating -based addition-subtration -hain for an integer n Input:,, n = l n i i = n 0,..., n l Output: -based addition-subtration -hain for n : vi i for 0 i l : j 0 3: for all x i 0 in x = {n 0, n,..., n l, n l } do 4: t j, t j+,..., t j+ xi i + signx i,..., 5: i + signx i 6: j j + x i 7: end for 8: t j, t j+,..., t j/ 0, 0,..., 0 9: for i = 0 to j/ do 0: vl + i + vl + i + vt i+ signt i+ : +vt i++ signt i++ : + 3: +vt i++ signt i++ 4: end for 5: return v There are some differenes between Algorithm and Algorithm 3. In Algorithm 3, the input is a base signed digit representation instead of traditional representation. We will use relaxed redued representation as an input here. In line 3-5 and 0-3, we use sign funtion to separate addition and subtration operations. We have to use + in line 3-5, use 0 in line 8 and use on line 0-3 to mae the index values not eual to zero. Example 6. We will follow Algorithm 3 to generate 4-based a/s 5-hain for n = 977. From Algorithm, the relaxed redued representation in base 4 for 977 is,, 0,, 0, 4 and we will use it as an input. Following Algorithm 3, in line, we will have v = [, 4, 6, 64, 56, 04] 5. By hoosing x i from left to right, at the end of line 7 we will have t =,,, 4, 6, 0, 0, 0. In line 9 to 4, the next element in the hain is alulated from the first 4 elements of t, v6 = v5 + v0 v v v3 = 953. Then, the last element is alulated from the last 4 elements of t, v7 = v6 + v5 + v + v + v = 977. The algorithm returns the 4-based a/s 5-hain for 977, [, 4, 6, 64, 56, 04, 953, 977] 5. We show the flow of the algorithm in Fig.. From Algorithms and 3, you an see that relaxed ost of the representation affets the length of the hain. In the next setion, we will disuss on the relaxed ost, the bound of the length of the hain and show proofs on this. These will lead to the omputation time of salar multipliation using -based a/s -hain tehniue. IV. WORST CASE OF COMPUTATION TIME USING -BASED ADDITION-SUBTRACTION -CHAIN TECHNIQUE Before showing the omputation time, we give some disussions on the relaxed redued representations and its maximum

5 = ε b = ε m > m Fig.. Generating 4-based a/s 5-hain of 977 using Algorithm 3 relaxed osts. We will use these relaxed redued representations as inputs for Algorithm 3. They also diretly affet the length of our -based a/s -hain and the omputation time. A. Worst Case of Relaxed Cost of Relaxed Redued Representation Lemmas and theorems disussed in Subsetion II-A show many properties of the relaxed redued representations and we will use these properties to prove our wors. We then onlude the worst ase of relaxed ost of relaxed redued representation in Theorem. Lemma 7. For all natural number m and odd base 3, there exists an integer 0 w m with its relaxed redued representation ε R w suh that ε = + m. Proof. Let b := and w := m +. The relaxed redued representation of w is ε = ε 0,..., ε m = b,..., b, R w. You an see that 0 w = m ε i i m and ε = + bm = + m. Lemma 8. For all natural number m and odd base 3, ε is at most + m for relaxed redued representation ε R w, 0 w m. Proof. Let b :=. We assume that there exists an integer 0 w m with a relaxed redued representation ε = ε 0,..., ε R w suh that ε > + m = + bm. Case the length > m, from relaxed redued representation s properties in [] that ε i b, we an derive the ineualities as follows. From the assumption, we now that ε > 0. w = ε i i = ε + ε i i ε + b i and this ontradits the fat that w m. Case the length < m, relaxed redued representation s properties in [] show that ε i b, so it is impossible to have < m with ε > + bm. Case the length = m and = 3, the relaxed ost of relaxed redued representation is maximized when w = m + whih has relaxed redued representation ε = ε 0,..., ε =,...,,. We have ε = + bm and this is a ontradition. Case the length = m and 3, if ε =, we have ε i + bm. From ε i b, this is impossible. Thus, the lowest ε we an have is and we get lowest w with relaxed ost of its relaxed redued representation greater than + bm is w = 3m + with its relaxed redued representation ε 0,..., ε = b,..., b, R w. You an see that w > m and this is a ontradition. Lemma 9. For all natural number m and even base, there exists an integer 0 w m with its relaxed redued representation ε R w suh that { + m if m is odd ε = + m if m is even Proof. Let b := and a :=. Case m is odd, let w odd := m a m a + b m with its representation ε = ε 0,..., ε m = a, b,..., a, b, a, R w odd. The redution rules in Definition 6 an be applied to show that ε is relaxed redued. You an see that 0 w odd = m ε i i m and ε = + a + a + b m = + m. Case m is even, let w even := m m + with its representation ε = ε 0,..., ε m = b, a,..., b, a, R w even. The redution rules in Definition 6 an be applied to show that ε is relaxed redued. You an see that 0 w even = m ε i i m and ε = +a+b m = + m. Lemma 0. For all natural number m and even base, ε is at most { ε + m if m is odd + m if m is even for relaxed redued representation ε R w, 0 w m. Proof. Let b := and a :=.

6 Case m is odd, we assume that there exists an integer 0 w m with its relaxed redued representation ε = ε 0,..., ε R w suh that ε > + m Case the length > m, using the argument similar to Lemma 8 for the ase when the length > m and some relaxed redued representation s properties, we have ε i i > m. We have a ontradition. Case the length < m, relaxed redued representation s properties in [] show that ε i b and there should be no two adjaent digits eual to b or b, so it is impossible to have < m with ε > + m. Case the length = m and =, using the properties on adjaent digits and the argument similar to Lemma 8 for the ase when the length = m and = 3, we will have a ontradition. Case the length = m and, using the argument similar to Lemma 8 for the ase when the length = m and 3, the lowest w we an have with relaxed ost of its relaxed redued representation greater than + m is w = 3 m + m + m + with relaxed redued representation ε 0,..., ε = a, b,..., a, b, a, R w. You an see that w > m and this is a ontradition. Case m is even, the proof an be ompleted using the similar argument. Finally, we onlude all lemmas in this theorem. Theorem maximum relaxed ost of relaxed redued representation. Given a natural number m and base, the maximum relaxed ost of relaxed redued representation ε R w for 0 w m, ε, an be desribed as follows. + ε m if is odd + m if is even and m is odd + m if is even and m is even B. Worst Case of Computation Time Using -Based Addition- Subtration -Chain Tehniue From Theorems 0, and Algorithm 3, we an propose a bound of the length of -based a/s -hain as follows. Theorem bound of the length of -based addition-subtration -hain. Given positive integer, and n, the bound of the optimal length s of -based addition-subtration -hain for an integer n an be desribed as follows. log n s log n + log n Note that log n is the length of the power part and log n is the length of the addition part. This an be alulated straightforward from Theorem 0 and Theorem. This theorem leads to the worst ase of the omputation time using -based a/s -hain tehniue with no parallel omputation. Refer to Subsetion II-C, the osts from using -based a/s -hain of n to alulate np are written in the following notation. We define m := log n + as number of bits for n. Avg m,, = Q m log + A avg m log +O, Max m,, = Q m m log + A max log +O, where Q is the ost of multiplying the point by. The onstants avg and max are the average and maximum relaxed osts for bit representation in base whih are explained in this setion and Subsetion II-A. V. PARALLEL SCALAR MULTIPLICATION USING -BASED ADDITION-SUBTRACTION -CHAIN TECHNIQUE In [], -hain is used to improve the inversion operation of an element in GF m. In this paper, we will use the proposed -based a/s -hain to improve salar multipliation in parallel way. We first introdue the most basi tehniue with no parallel omputation, namely -based a/s -hain tehniue. We then propose a -based a/s 4-hain parallel tehniue whih leads to our main ontribution, -based a/s - hain parallel tehniue. We also disuss an optimization for multiple operand point addition using hybrid-double multiplier [9], [0]. A. -Based Addition-Subtration 4-Chain Parallel Tehniue We start by introduing -based a/s -hain tehniue whih is the most simple tehniue. We an alulate np by following the way -based a/s -hain of n is onstruted. We alulate point doubling in power part and point addition in the addition part of the hain. From Theorem 8, Theorem, Algorithms and Algorithm 3, the length of the power part is at most m := log n +. The length of the addition part is at average 3 m + O and at most m + O. Thus, the average and maximum osts for -based a/s -hain tehniue an be written with the following notations. Avg m,, = D m + A 3 m + O, Max m,, = D m + A m + O. This is the most simple tehniue using -based a/s -hain. You may notie that using NAF tehniue will have the same ost. Next, we will move to the parallel tehniue by using parallel point addition. We start with -based a/s 4 hain, and then generalize to -based a/s hain in the next subsetion. The idea is that, to alulate P + P + P 3 + P 4, parallel point addition will firstly do P = P + P and P = P 3 + P 4 simultaneously with omputing ores. Then the final answer is P +P. Thus, this osts only A instead of 3A. We define 4-operand parallel point addition A whih euals to A.

7 Remar: Given ellipti points P i and positive integer x, x- operand point addition is the alulation of P +P + +P x. To use -based a/s 4-hain parallel tehniue to alulate np, we do point doubling in power part and then do 4-operand parallel point addition in addition part. The length of the power part is at most m := log n +. By Theorems 8 and, the length of the addition part is 3 3 m + O at average and, by Theorems and, at most 3 m + O. Thus, the average and maximum osts for -based a/s 4-hain parallel point addition tehniue with omputing ores an be written with the following notations. Avg m,,4, = D m + A 9 m + O = D m + A 9 m + O, Max m,,4, = D m + A 6 m + O = D m + A 3 m + O. Another easy way to alulate the ost is to multiply a fator of 3 from the parallel point addition to the addition part of -based a/s -hain. B. Generalize to -Based Addition-Subtration -Chain Tehniue with Computing Cores If you have enough omputing resoures, this tehniue an be generalized by using -based a/s -hain with omputing ores an exeute point additions at the same time. Start with operands points, beause it is a -hain, we assume that number of the operands are redued by half eah round with -operand addition, thus we have log rounds. The number of point additions in eah round is the number of the operands remaining in eah round divides by times the number of omputing ores, rounding up. Similar to the last subsetion, the average and maximum osts for the generalized tehniue an be written as follows. Avg m,odd,, = Q m log + A log Avg m,even,, = Q m Max m,,, = Q m + 4 log + A log + A m log + O log m log + O m log + O log Another easy way to approximate the ost is to multiply a fator of from the parallel omputing ores to the addition part. Remar: To get the minimum ost, we found that we should use = -based and have the highest as possible. The hoies for have only small effet. We will disuss on this topi in the next setion. Moreover, using = is also effiient in GF m beause the doubling operation is fast. The value of an be hosen depending on the finite field used. C. Optimizing Multiple Operand Point Addition for Twisted Edwards Curves In this subsetion, we disuss an optimization tehniue for -based a/s -hain. Although we an use any for -based a/s -hain tehniue, the last subsetion, Subsetion V-B, show that = will lead to the minimum ost and we will use this value through this subsetion as an example on how to apply the optimization. Other values of are also possible for this tehniue. Firstly, we give some explanations about the finite field elements multipliation. The -operand finite field elements multipliation is an operation that an give a multipliation result of two field elements, while the 3-operand finite field elements multipliation is an operation that an give a multipliation result of three field elements. For example, when λ, λ, λ 3 are elements of a field, the -operand finite field elements multipliation an give a value of λ λ, while 3- operand finite field elements multipliation an give a value of λ λ λ 3. In [9], [0], a iruit for the 3-operand finite field elements multipliation alled hybrid-double multiplier is given. The alulation time of the iruit is almost eual to that of the fastest -operand finite field elements multipliation. Remar: The finite field elements multipliation is a part of the ellipti point addition. They are not the same operations. We will fous on twisted Edwards urves [8], ax + y = + dx y, beause they are the urrently fastest urves []. Cost for -operand point addition an operation that alulate P + P when P and P are points on an ellipti urve using hybrid-double multiplier is A = 9M. We define A as the ost for 3-operand point addition an operation that alulate P + P + P 3 when P, P, and P 3 are points on an ellipti urve. We an rearrange some part of the point addition euations from [8] and have 7M as the ost for A. The rearranged euations for A are explained as follows. Let X i, Y i and Z i be finite field elements. Given ellipti points P X, Y, Z, P X, Y, Z and P 3 X 3, Y 3, Z 3 in projetive oordinate where the x-y oordinate x i, y i = X i /Z i, Y i /Z i. We use most euations from [8] beause they are the urrently fastest formulas reorded in []. First, we alulate X and Y from P and P. λ = Z Z λ = λ λ 3 = X X λ 4 = Y Y λ 5 = dλ 3 λ 4 λ 6 = λ λ 5 λ 7 = λ + λ 5

8 X = λ λ 6 X + Y X + Y λ 3 λ 4 Y = λ λ 7 λ 4 aλ 3 The alulation ost is now 8M using hybrid-double multiplier. Notie that we do not alulate Z here. Calulate Z = λ 6 λ 7 will finish the -operand point addition with the ost of 9M. We ontinue to alulate P 4 X 4, Y 4, Z 4 = P + P + P 3 as follows and you an see that the total ost is 7M. λ 8 = λ 6 λ 7 Z 3 λ 9 = λ 8 λ 0 = X X 3 λ = Y Y 3 λ = dλ 0 λ λ 3 = λ 9 λ λ 4 = λ 9 + λ X 4 = λ 8 λ 3 X + X 3 Y + Y 3 λ 0 λ Y 4 = λ 8 λ 4 λ aλ 0 Z 4 = λ 3 λ 4 It is important to note that with the hybrid-double multiplier, we have 7M = A < A = 8M. The lower ost omes from the λ 8 euation. To alulate np from -based a/s - hain, we do point doubling in the power part and then do 3-operand point addition in the addition part. In this way, the average and maximum osts for -based a/s -hain tehniue with ores are Avg m,optimize,, = D m + A 3 log 3 7 m + O 54 3 log 3 Max m,optimize,, = D m + A 7 36 m + O Another easy way to alulate the ost is to multiply a fator of 7 8 from the 3-operand point addition to the addition part. We an generalize this optimization to x-operand point addition for any positive integer x 3. Calulating the summation of x ellipti points with hybrid-double multiplier but no multiple operand point addition will ost 9x M. For eah operand higher than, if we use the hybriddouble multiplier, we an redue the finite field multipliation by one. This is beause we do not have to alulate the intermediate value of Z i. Refer to the rearrange euations earlier, you an see that we an alulate λ 8 = λ 6 λ 7 Z 3 with only one 3-operand field elements multipliation instead of Z = λ 6 λ 7 and λ 8 = Z Z 3 separately. You an apply this 3-operand field element multipliation to every point operand added. This means that, using x-operand point addition will redue x M from the ost. This will give a fator of to the addition part. 9x x 9x = 8x 7 9x 9 Reall that the number of points we want to add using ores is from the property of -based a/s -hain. We now that the omputation time is optimized when eah ore has almost the same number of points to add. Beause of that, eah ore has the point to alulate the summation at most. Having x > will not give any more improvements. Thus, the value of x that minimize our ost is. If we an rearrange the euations in other ways and redue more field element multipliations, we an similarly apply the tehniue to lower the alulation ost. However, we found during the ourse of this wor that the further redution is very diffiult. D. Side Channel Atta Prevention In previous tehniues inluding double-and-add, NAF, w- NAF, addition hain and addition-subtration hain, we an now the seuene of addition and doubling steps by monitoring the power power atta [] or time used timing atta [3]. These inds of attas are alled side-hannel atta. Some tehniues, inluding Montgomery ladder [4], deploy dummy operations or use uniform operations to enounter this, however, the alulation osts beome worse. In -based a/s -hain tehniues, all alulations begin with some numbers of multiplying and then follow with some additions. The n in np annot be distinguished using these ind of seuenes. Thus, we do not need any dummy operations. This is how we resist power atta. For timing atta in -based a/s -hain, one an measure the time used in the addition part and tell how many times the addition are exeuted. Some n may have short addition part, thus, it is easy to guess this integer n. We an add some dummy operations and mae all exeutions eual to the maximum omputation time. Even though the alulation ost beomes worse, we have proved that the maximum omputation time is still effiient. In this way, we an prevent timing atta and also power atta. VI. COMPARISONS AND PARAMETER SETTINGS A. Comparisons with Tehniues without Windows Defining m := log n +, from Setions V, the osts Avg m and Max m of all tehniues using base representations are in the form D m + A oeff. m + O. Table I summarizes the oeffiient values of Am for average and worst ase. w-naf and wr-naf are not inluded in the table beause their alulation osts are affeted from the preomputation osts. Note that the multiple operand point addition optimization is now speifi to twisted Edwards urves due to the rearranged euations. The other tehniues an be generally applied to any urves. Aording to Table I, among the proposed tehniue, hoosing -based a/s -hain with large number of ores,, and a multiple operand point addition optimization will lead to the minimum ost.

9 TABLE I COEFFICIENT VALUE OF Am FOR EACH TECHNIQUE Tehniue used Average oeffiient Maximum oeffiient Double-and-add -based a/s -hain / NAF [] -based a/s 4-hain ores -based a/s -hain ores log = = log based a/s -hain ores with 3-operand addition and hybrid-double multiplier -based a/s -hain ores with x-operand addition and hybrid-double multiplier 3 log x log x 8x x log x log x 8x 7 4 9x B. Comparisons with w-naf and Montgomery Ladder We first preisely ompare the maximum ost between w-naf representation and our -based a/s -hain parallel tehniue. Beause of the pre-omputation ost in w-naf, we have to speify the number of bits used to represent an integer. We hoose 8 and 56 bit integers for this purpose []. The maximum ost should be onsidered in this ase due to the aims of side-hannel atta. For w-naf wr-naf with r =, [4] suggests the best w to be 5 and we will use these values in this subsetion. Aording to Subsetions II-C and V-B, the maximum ost for 8 bits would be Max 8,5-NAF = D 8 + A = 8 D A, Max 8,,, = D 8 + A log 8 + O In this ase, using -based a/s -hain tehniue with only omputing ores will have slightly smaller ost. You an improve the result by using more than omputing ores. Next, the maximum ost for 56 bits would be Max 56,5-NAF = D 56 + A = 56 D A, Max 56,,, = D 56 + A log 56 + O In 56-bit ase, using -based a/s hain tehniue with 3 omputing ores or more will have smaller ost. Approximately, for large m, our hain tehniue with 3 omputing ores will have up to /5 4/7 /5 = 5.9% less additions. but if using omputing ores, our tehniue an have up to 4/8 /5 /5 =.% more additions. Similarly, in base 3 representation and GF 3 m, [4] suggests w = 3 and r = 3 for wr-naf tehniue. Our 3-based a/s -hain tehniue an overome the urrent ost using only 3 omputing ores with 3-operand addition optimization and hybrid-double multiplier. We show the approximated osts as follows and use T as the ost for tripling. Max m,3,3-naf = T m log 3 + A m 3 log 3 +O Max m,opt3,, = T m 7 m log 3 + A 8 log 3 +O For Montgomery ladder [4], the tehniue uses addition for eah digit in the binary representation. Using the number of bits with the same notation m = log n +, the maximum ost for alulating np using this tehniue is Max m,mont = A m Beause the ost of addition A = M is more than doubling D = 7M with no hybrid-double multiplier, our -based a/s -hain tehniues an perform better with omputing ores and above. If the power part is already preomputed, our hain tehniue an perform better in any ase. C. Choie of in -Based Addition-Subtration -Chain Although, hoies for do not have muh effets on the ost, you an see that hoosing large for -based a/s -hain an lower the alulation ost owing to the ost euation, the

10 rounding up and the optimization tehniue. However, using smaller also has some benefits. In ellipti urve ryptography, we now P long before nowing n. The power part an be alulated before n is given. When n is given, it will go through Algorithm to generate its relaxed redued representation. After the algorithm produes some least signifiant digits, if we hoose small, we an alulate the next element of the hain immediately, not having to wait for the algorithm to finish. The smaller we hoose, the earlier we an start the alulation. Smaller also has another benefit. Some appliations need the results of np for the same P and some different n. Using smaller, some ommon digits between the different n an be alulated only one. See the example below. Example 3. To alulate 45P and 77P from -based a/s -hain, we have relaxed redued representation for 45 =, 0,, 0,, 0, and 77 =, 0,, 0,, 0,. From Algorithm 3, we an hoose x i eah digit in relaxed redued representation in order to have t for 45 =, 3, 5, 0 and t for 77 =, 3, 5, 0. If we hoose = 3 - based a/s 3-hain, we have [,, 4, 8, 6, 3, 64, 6, 45] for 45 and [,, 4, 8, 6, 3, 64, 6, 77] for 77. Notie that we an alulate the ommon 6P one. We an redue more addition if n is large. In this example, the tehniue annot be used if we hoose 4. From our experiments, we random 00,000 pairs of 8-bit integers and generate their base relaxed redued representations. At average, there are 3 ommon digits between eah pair. Changing to base 3 an inrease this value to 7. This means if we hoose the appropriate value for, at average, we an redue 3 and 7 additions, respetively. We may use this tehniue and onsider more than integers at the same time. A method for seleting whih digit to alulate, its effiieny and method to find the optimized are left for further studies. VII. CONCLUSIONS AND FUTURE WORKS Parallel tehniues have been used widely for effiient omputations, but not many parallel tehniues have been proposed to improve salar multipliation with side-hannel atta prevention. Our paper propose a parallel salar multipliation tehniue using -based a/s -hain. The proposed tehniue uses the omputing ores effiiently as we an see that using omputing ores an redue the addition osts to the fator of. This parallel tehniues an perform better than the urrent w-naf tehniue by using only 3 ores. We also have some optimization of multiple operand point addition whih an redue the addition osts up to the fator of 8 9. At the end, we show that our parallel -based a/s -hain tehniue an resist side-hannel atta effiiently. Some ellipti urve multiple operand addition euations rearrangements, method for seleting digits and the optimized value for are left for further studies. ACKNOWLEDGMENT This researh was mainly done while Kittiphop Phalaarn and Kittiphon Phalaarn did an internship at The University of Toyo. The authors would lie to than The University of Toyo, Assis. Prof. Athasit Surarers and Prof. Hiroshi Imai for failitating the internship. REFERENCES [] D. Hanerson, A. J. Menezes, and S. Vanstone, Guide to ellipti urve ryptography. Springer Siene & Business Media, 006. [] W. Bosma, Signed bits and fast exponentiation, Journal de théorie des nombres de Bordeaux, vol. 3, no., pp. 7 4, 00. [3] K. Oeya and T. Taagi, The width-w NAF method provides small memory and fast ellipti salar multipliations seure against side hannel attas, in Cryptographers Tra at the RSA Conferene, pp , Springer, 003. [4] T. Taagi, D. J. Reis, S.-M. Yen, and B.-C. Wu, Radix-r non-adjaent form and its appliation to pairing-based ryptosystem, IEICE Transations on Fundamentals of Eletronis, Communiations and Computer Sienes, vol. 89, no., pp. 5 3, 006. [5] D. E. Knuth, The Art of Computer Programming, Volume : Seminumerial Algorithms. Addison-Wesley, 997. [6] H. Volger, Some results on addition/subtration hains, Information Proessing Letters, vol. 0, no. 3, pp , 985. [7] N. Meloni, Fast and seure ellipti urve salar multipliation over prime fields using speial addition hains, IACR Cryptology eprint Arhive, vol , 006. [8] R. R. Goundar, K. Shiota, and M. Toyonaga, New strategy for doublingfree short addition-subtration hain, International Journal of Applied Mathematis, vol., no. 3, pp , 007. [9] H. M. Bahig, On a generalization of addition hains: Addition multipliation hains, Disrete Mathematis, vol. 308, no. 4, pp. 6 66, 008. [0] M. Nöer, Some remars on parallel exponentiation, in International Symposium on Symboli and Algebrai Computation, pp , ACM, 000. [] K. Järvinen, V. Dimitrov, and R. Azarderahsh, A generalization of addition hains and fast inversions in binary fields, IEEE Transations on Computers, vol. 64, no. 9, pp. 4 43, 05. [] P. Koher, J. Jaffe, and B. Jun, Differential power analysis, in Annual International Cryptology Conferene, pp , Springer, 999. [3] P. C. Koher, Timing attas on implementations of Diffie-Hellman, RSA, DSS, and other systems, in Annual International Cryptology Conferene, pp. 04 3, Springer, 996. [4] M. Joye and S.-M. Yen, The Montgomery powering ladder, in International Worshop on Cryptographi Hardware and Embedded Systems, pp. 9 30, Springer, 00. [5] N. Méloni and M. A. Hasan, Random digit representation of integers, in IEEE Symposium on Computer Arithmeti in press, 06. [6] B. Möller, Seuring ellipti urve point multipliation against sidehannel attas, in International Conferene on Information Seurity, pp , Springer, 00. [7] C. Heuberger and H. Prodinger, On minimal expansions in redundant number systems: Algorithms and uantitative analysis, Computing, vol. 66, no. 4, pp , 00. [8] D. J. Bernstein, P. Birner, M. Joye, T. Lange, and C. Peters, Twisted Edwards urves, in International Conferene on Cryptology in Afria AfriaCrypt 08, pp , Springer, 008. [9] R. Azarderahsh and A. Reyhani-Masoleh, Low-omplexity multiplier arhitetures for single and hybrid-double multipliations in Gaussian normal bases, IEEE Transations on Computers, vol. 6, no. 4, pp , 03. [0] R. Azarderahsh and A. Reyhani-Masoleh, Parallel and high-speed omputations of ellipti urve ryptography using hybrid-double multipliers, IEEE Transations on Parallel and Distributed Systems, vol. 6, no. 6, pp , 05. [] Expliit-formulas database. Aessed July, 06. [] A. Jivsov, Ellipti urve ryptography ECC in OpenPGP, 0.

The Effectiveness of the Linear Hull Effect

The Effectiveness of the Linear Hull Effect The Effetiveness of the Linear Hull Effet S. Murphy Tehnial Report RHUL MA 009 9 6 Otober 009 Department of Mathematis Royal Holloway, University of London Egham, Surrey TW0 0EX, England http://www.rhul.a.uk/mathematis/tehreports

More information

Millennium Relativity Acceleration Composition. The Relativistic Relationship between Acceleration and Uniform Motion

Millennium Relativity Acceleration Composition. The Relativistic Relationship between Acceleration and Uniform Motion Millennium Relativity Aeleration Composition he Relativisti Relationship between Aeleration and niform Motion Copyright 003 Joseph A. Rybzyk Abstrat he relativisti priniples developed throughout the six

More information

Lightpath routing for maximum reliability in optical mesh networks

Lightpath routing for maximum reliability in optical mesh networks Vol. 7, No. 5 / May 2008 / JOURNAL OF OPTICAL NETWORKING 449 Lightpath routing for maximum reliability in optial mesh networks Shengli Yuan, 1, * Saket Varma, 2 and Jason P. Jue 2 1 Department of Computer

More information

The Hanging Chain. John McCuan. January 19, 2006

The Hanging Chain. John McCuan. January 19, 2006 The Hanging Chain John MCuan January 19, 2006 1 Introdution We onsider a hain of length L attahed to two points (a, u a and (b, u b in the plane. It is assumed that the hain hangs in the plane under a

More information

Maximum Entropy and Exponential Families

Maximum Entropy and Exponential Families Maximum Entropy and Exponential Families April 9, 209 Abstrat The goal of this note is to derive the exponential form of probability distribution from more basi onsiderations, in partiular Entropy. It

More information

23.1 Tuning controllers, in the large view Quoting from Section 16.7:

23.1 Tuning controllers, in the large view Quoting from Section 16.7: Lesson 23. Tuning a real ontroller - modeling, proess identifiation, fine tuning 23.0 Context We have learned to view proesses as dynami systems, taking are to identify their input, intermediate, and output

More information

arxiv:math/ v4 [math.ca] 29 Jul 2006

arxiv:math/ v4 [math.ca] 29 Jul 2006 arxiv:math/0109v4 [math.ca] 9 Jul 006 Contiguous relations of hypergeometri series Raimundas Vidūnas University of Amsterdam Abstrat The 15 Gauss ontiguous relations for F 1 hypergeometri series imply

More information

A NETWORK SIMPLEX ALGORITHM FOR THE MINIMUM COST-BENEFIT NETWORK FLOW PROBLEM

A NETWORK SIMPLEX ALGORITHM FOR THE MINIMUM COST-BENEFIT NETWORK FLOW PROBLEM NETWORK SIMPLEX LGORITHM FOR THE MINIMUM COST-BENEFIT NETWORK FLOW PROBLEM Cen Çalışan, Utah Valley University, 800 W. University Parway, Orem, UT 84058, 801-863-6487, en.alisan@uvu.edu BSTRCT The minimum

More information

Sensitivity Analysis in Markov Networks

Sensitivity Analysis in Markov Networks Sensitivity Analysis in Markov Networks Hei Chan and Adnan Darwihe Computer Siene Department University of California, Los Angeles Los Angeles, CA 90095 {hei,darwihe}@s.ula.edu Abstrat This paper explores

More information

General Equilibrium. What happens to cause a reaction to come to equilibrium?

General Equilibrium. What happens to cause a reaction to come to equilibrium? General Equilibrium Chemial Equilibrium Most hemial reations that are enountered are reversible. In other words, they go fairly easily in either the forward or reverse diretions. The thing to remember

More information

Complexity of Regularization RBF Networks

Complexity of Regularization RBF Networks Complexity of Regularization RBF Networks Mark A Kon Department of Mathematis and Statistis Boston University Boston, MA 02215 mkon@buedu Leszek Plaskota Institute of Applied Mathematis University of Warsaw

More information

Journal of Inequalities in Pure and Applied Mathematics

Journal of Inequalities in Pure and Applied Mathematics Journal of Inequalities in Pure and Applied Mathematis A NEW ARRANGEMENT INEQUALITY MOHAMMAD JAVAHERI University of Oregon Department of Mathematis Fenton Hall, Eugene, OR 97403. EMail: javaheri@uoregon.edu

More information

THE NUMBER FIELD SIEVE FOR INTEGERS OF LOW WEIGHT

THE NUMBER FIELD SIEVE FOR INTEGERS OF LOW WEIGHT MATHEMATICS OF COMPUTATION Volume 79, Number 269, January 2010, Pages 583 602 S 0025-5718(09)02198-X Artile eletronially published on July 27, 2009 THE NUMBER FIELD SIEVE FOR INTEGERS OF LOW WEIGHT OLIVER

More information

max min z i i=1 x j k s.t. j=1 x j j:i T j

max min z i i=1 x j k s.t. j=1 x j j:i T j AM 221: Advaned Optimization Spring 2016 Prof. Yaron Singer Leture 22 April 18th 1 Overview In this leture, we will study the pipage rounding tehnique whih is a deterministi rounding proedure that an be

More information

Product Policy in Markets with Word-of-Mouth Communication. Technical Appendix

Product Policy in Markets with Word-of-Mouth Communication. Technical Appendix rodut oliy in Markets with Word-of-Mouth Communiation Tehnial Appendix August 05 Miro-Model for Inreasing Awareness In the paper, we make the assumption that awareness is inreasing in ustomer type. I.e.,

More information

Counting Idempotent Relations

Counting Idempotent Relations Counting Idempotent Relations Beriht-Nr. 2008-15 Florian Kammüller ISSN 1436-9915 2 Abstrat This artile introdues and motivates idempotent relations. It summarizes haraterizations of idempotents and their

More information

Wavetech, LLC. Ultrafast Pulses and GVD. John O Hara Created: Dec. 6, 2013

Wavetech, LLC. Ultrafast Pulses and GVD. John O Hara Created: Dec. 6, 2013 Ultrafast Pulses and GVD John O Hara Created: De. 6, 3 Introdution This doument overs the basi onepts of group veloity dispersion (GVD) and ultrafast pulse propagation in an optial fiber. Neessarily, it

More information

Lyapunov Exponents of Second Order Linear Systems

Lyapunov Exponents of Second Order Linear Systems Reent Researhes in Computational Intelligene and Information Seurity Lyapunov Exponents of Seond Order Linear Systems ADAM CZORNIK and ALEKSANDER NAWRAT Department of Automati Control Silesian Tehnial

More information

UPPER-TRUNCATED POWER LAW DISTRIBUTIONS

UPPER-TRUNCATED POWER LAW DISTRIBUTIONS Fratals, Vol. 9, No. (00) 09 World Sientifi Publishing Company UPPER-TRUNCATED POWER LAW DISTRIBUTIONS STEPHEN M. BURROUGHS and SARAH F. TEBBENS College of Marine Siene, University of South Florida, St.

More information

Word of Mass: The Relationship between Mass Media and Word-of-Mouth

Word of Mass: The Relationship between Mass Media and Word-of-Mouth Word of Mass: The Relationship between Mass Media and Word-of-Mouth Roman Chuhay Preliminary version Marh 6, 015 Abstrat This paper studies the optimal priing and advertising strategies of a firm in the

More information

Canimals. borrowed, with thanks, from Malaspina University College/Kwantlen University College

Canimals. borrowed, with thanks, from Malaspina University College/Kwantlen University College Canimals borrowed, with thanks, from Malaspina University College/Kwantlen University College http://ommons.wikimedia.org/wiki/file:ursus_maritimus_steve_amstrup.jpg Purpose Investigate the rate of heat

More information

SERIJA III

SERIJA III SERIJA III www.math.hr/glasnik I. Gaál, B. Jadrijević and L. Remete Totally real Thue inequalities over imaginary quadrati fields Aepted manusript This is a preliminary PDF of the author-produed manusript

More information

FINITE WORD LENGTH EFFECTS IN DSP

FINITE WORD LENGTH EFFECTS IN DSP FINITE WORD LENGTH EFFECTS IN DSP PREPARED BY GUIDED BY Snehal Gor Dr. Srianth T. ABSTRACT We now that omputers store numbers not with infinite preision but rather in some approximation that an be paed

More information

arxiv: v2 [math.pr] 9 Dec 2016

arxiv: v2 [math.pr] 9 Dec 2016 Omnithermal Perfet Simulation for Multi-server Queues Stephen B. Connor 3th Deember 206 arxiv:60.0602v2 [math.pr] 9 De 206 Abstrat A number of perfet simulation algorithms for multi-server First Come First

More information

Discrete Bessel functions and partial difference equations

Discrete Bessel functions and partial difference equations Disrete Bessel funtions and partial differene equations Antonín Slavík Charles University, Faulty of Mathematis and Physis, Sokolovská 83, 186 75 Praha 8, Czeh Republi E-mail: slavik@karlin.mff.uni.z Abstrat

More information

Nonreversibility of Multiple Unicast Networks

Nonreversibility of Multiple Unicast Networks Nonreversibility of Multiple Uniast Networks Randall Dougherty and Kenneth Zeger September 27, 2005 Abstrat We prove that for any finite direted ayli network, there exists a orresponding multiple uniast

More information

Geometry of Transformations of Random Variables

Geometry of Transformations of Random Variables Geometry of Transformations of Random Variables Univariate distributions We are interested in the problem of finding the distribution of Y = h(x) when the transformation h is one-to-one so that there is

More information

7 Max-Flow Problems. Business Computing and Operations Research 608

7 Max-Flow Problems. Business Computing and Operations Research 608 7 Max-Flow Problems Business Computing and Operations Researh 68 7. Max-Flow Problems In what follows, we onsider a somewhat modified problem onstellation Instead of osts of transmission, vetor now indiates

More information

Assessing the Performance of a BCI: A Task-Oriented Approach

Assessing the Performance of a BCI: A Task-Oriented Approach Assessing the Performane of a BCI: A Task-Oriented Approah B. Dal Seno, L. Mainardi 2, M. Matteui Department of Eletronis and Information, IIT-Unit, Politenio di Milano, Italy 2 Department of Bioengineering,

More information

Sufficient Conditions for a Flexible Manufacturing System to be Deadlocked

Sufficient Conditions for a Flexible Manufacturing System to be Deadlocked Paper 0, INT 0 Suffiient Conditions for a Flexile Manufaturing System to e Deadloked Paul E Deering, PhD Department of Engineering Tehnology and Management Ohio University deering@ohioedu Astrat In reent

More information

MOLECULAR ORBITAL THEORY- PART I

MOLECULAR ORBITAL THEORY- PART I 5.6 Physial Chemistry Leture #24-25 MOLECULAR ORBITAL THEORY- PART I At this point, we have nearly ompleted our rash-ourse introdution to quantum mehanis and we re finally ready to deal with moleules.

More information

Appendix A Market-Power Model of Business Groups. Robert C. Feenstra Deng-Shing Huang Gary G. Hamilton Revised, November 2001

Appendix A Market-Power Model of Business Groups. Robert C. Feenstra Deng-Shing Huang Gary G. Hamilton Revised, November 2001 Appendix A Market-Power Model of Business Groups Roert C. Feenstra Deng-Shing Huang Gary G. Hamilton Revised, Novemer 200 Journal of Eonomi Behavior and Organization, 5, 2003, 459-485. To solve for the

More information

Panos Kouvelis Olin School of Business Washington University

Panos Kouvelis Olin School of Business Washington University Quality-Based Cometition, Profitability, and Variable Costs Chester Chambers Co Shool of Business Dallas, TX 7575 hamber@mailosmuedu -768-35 Panos Kouvelis Olin Shool of Business Washington University

More information

A variant of Coppersmith s Algorithm with Improved Complexity and Efficient Exhaustive Search

A variant of Coppersmith s Algorithm with Improved Complexity and Efficient Exhaustive Search A variant of Coppersmith s Algorithm with Improved Complexity and Effiient Exhaustive Searh Jean-Sébastien Coron 1, Jean-Charles Faugère 2, Guénaël Renault 2, and Rina Zeitoun 2,3 1 University of Luxembourg

More information

Where as discussed previously we interpret solutions to this partial differential equation in the weak sense: b

Where as discussed previously we interpret solutions to this partial differential equation in the weak sense: b Consider the pure initial value problem for a homogeneous system of onservation laws with no soure terms in one spae dimension: Where as disussed previously we interpret solutions to this partial differential

More information

Verka Prolović Chair of Civil Engineering Geotechnics, Faculty of Civil Engineering and Architecture, Niš, R. Serbia

Verka Prolović Chair of Civil Engineering Geotechnics, Faculty of Civil Engineering and Architecture, Niš, R. Serbia 3 r d International Conferene on New Developments in Soil Mehanis and Geotehnial Engineering, 8-30 June 01, Near East University, Niosia, North Cyprus Values of of partial fators for for EC EC 7 7 slope

More information

Developing Excel Macros for Solving Heat Diffusion Problems

Developing Excel Macros for Solving Heat Diffusion Problems Session 50 Developing Exel Maros for Solving Heat Diffusion Problems N. N. Sarker and M. A. Ketkar Department of Engineering Tehnology Prairie View A&M University Prairie View, TX 77446 Abstrat This paper

More information

The First Integral Method for Solving a System of Nonlinear Partial Differential Equations

The First Integral Method for Solving a System of Nonlinear Partial Differential Equations ISSN 179-889 (print), 179-897 (online) International Journal of Nonlinear Siene Vol.5(008) No.,pp.111-119 The First Integral Method for Solving a System of Nonlinear Partial Differential Equations Ahmed

More information

Taste for variety and optimum product diversity in an open economy

Taste for variety and optimum product diversity in an open economy Taste for variety and optimum produt diversity in an open eonomy Javier Coto-Martínez City University Paul Levine University of Surrey Otober 0, 005 María D.C. Garía-Alonso University of Kent Abstrat We

More information

Packing Plane Spanning Trees into a Point Set

Packing Plane Spanning Trees into a Point Set Paking Plane Spanning Trees into a Point Set Ahmad Biniaz Alfredo Garía Abstrat Let P be a set of n points in the plane in general position. We show that at least n/3 plane spanning trees an be paked into

More information

Control Theory association of mathematics and engineering

Control Theory association of mathematics and engineering Control Theory assoiation of mathematis and engineering Wojieh Mitkowski Krzysztof Oprzedkiewiz Department of Automatis AGH Univ. of Siene & Tehnology, Craow, Poland, Abstrat In this paper a methodology

More information

Reliability Guaranteed Energy-Aware Frame-Based Task Set Execution Strategy for Hard Real-Time Systems

Reliability Guaranteed Energy-Aware Frame-Based Task Set Execution Strategy for Hard Real-Time Systems Reliability Guaranteed Energy-Aware Frame-Based ask Set Exeution Strategy for Hard Real-ime Systems Zheng Li a, Li Wang a, Shuhui Li a, Shangping Ren a, Gang Quan b a Illinois Institute of ehnology, Chiago,

More information

Generalized Dimensional Analysis

Generalized Dimensional Analysis #HUTP-92/A036 7/92 Generalized Dimensional Analysis arxiv:hep-ph/9207278v1 31 Jul 1992 Howard Georgi Lyman Laboratory of Physis Harvard University Cambridge, MA 02138 Abstrat I desribe a version of so-alled

More information

Danielle Maddix AA238 Final Project December 9, 2016

Danielle Maddix AA238 Final Project December 9, 2016 Struture and Parameter Learning in Bayesian Networks with Appliations to Prediting Breast Caner Tumor Malignany in a Lower Dimension Feature Spae Danielle Maddix AA238 Final Projet Deember 9, 2016 Abstrat

More information

A Characterization of Wavelet Convergence in Sobolev Spaces

A Characterization of Wavelet Convergence in Sobolev Spaces A Charaterization of Wavelet Convergene in Sobolev Spaes Mark A. Kon 1 oston University Louise Arakelian Raphael Howard University Dediated to Prof. Robert Carroll on the oasion of his 70th birthday. Abstrat

More information

7.1 Roots of a Polynomial

7.1 Roots of a Polynomial 7.1 Roots of a Polynomial A. Purpose Given the oeffiients a i of a polynomial of degree n = NDEG > 0, a 1 z n + a 2 z n 1 +... + a n z + a n+1 with a 1 0, this subroutine omputes the NDEG roots of the

More information

FORMAL METHODS LECTURE VI BINARY DECISION DIAGRAMS (BDD S)

FORMAL METHODS LECTURE VI BINARY DECISION DIAGRAMS (BDD S) Alessandro Artale (FM First Semester 2009/2010) p. 1/38 FORMAL METHODS LECTURE VI BINARY DECISION DIAGRAMS (BDD S) Alessandro Artale Faulty of Computer Siene Free University of Bolzano artale@inf.unibz.it

More information

Advances in Radio Science

Advances in Radio Science Advanes in adio Siene 2003) 1: 99 104 Copernius GmbH 2003 Advanes in adio Siene A hybrid method ombining the FDTD and a time domain boundary-integral equation marhing-on-in-time algorithm A Beker and V

More information

Subject: Introduction to Component Matching and Off-Design Operation % % ( (1) R T % (

Subject: Introduction to Component Matching and Off-Design Operation % % ( (1) R T % ( 16.50 Leture 0 Subjet: Introdution to Component Mathing and Off-Design Operation At this point it is well to reflet on whih of the many parameters we have introdued (like M, τ, τ t, ϑ t, f, et.) are free

More information

CMSC 451: Lecture 9 Greedy Approximation: Set Cover Thursday, Sep 28, 2017

CMSC 451: Lecture 9 Greedy Approximation: Set Cover Thursday, Sep 28, 2017 CMSC 451: Leture 9 Greedy Approximation: Set Cover Thursday, Sep 28, 2017 Reading: Chapt 11 of KT and Set 54 of DPV Set Cover: An important lass of optimization problems involves overing a ertain domain,

More information

A Spatiotemporal Approach to Passive Sound Source Localization

A Spatiotemporal Approach to Passive Sound Source Localization A Spatiotemporal Approah Passive Sound Soure Loalization Pasi Pertilä, Mikko Parviainen, Teemu Korhonen and Ari Visa Institute of Signal Proessing Tampere University of Tehnology, P.O.Box 553, FIN-330,

More information

DIGITAL DISTANCE RELAYING SCHEME FOR PARALLEL TRANSMISSION LINES DURING INTER-CIRCUIT FAULTS

DIGITAL DISTANCE RELAYING SCHEME FOR PARALLEL TRANSMISSION LINES DURING INTER-CIRCUIT FAULTS CHAPTER 4 DIGITAL DISTANCE RELAYING SCHEME FOR PARALLEL TRANSMISSION LINES DURING INTER-CIRCUIT FAULTS 4.1 INTRODUCTION Around the world, environmental and ost onsiousness are foring utilities to install

More information

LECTURE NOTES FOR , FALL 2004

LECTURE NOTES FOR , FALL 2004 LECTURE NOTES FOR 18.155, FALL 2004 83 12. Cone support and wavefront set In disussing the singular support of a tempered distibution above, notie that singsupp(u) = only implies that u C (R n ), not as

More information

General Closed-form Analytical Expressions of Air-gap Inductances for Surfacemounted Permanent Magnet and Induction Machines

General Closed-form Analytical Expressions of Air-gap Inductances for Surfacemounted Permanent Magnet and Induction Machines General Closed-form Analytial Expressions of Air-gap Indutanes for Surfaemounted Permanent Magnet and Indution Mahines Ronghai Qu, Member, IEEE Eletroni & Photoni Systems Tehnologies General Eletri Company

More information

REFINED UPPER BOUNDS FOR THE LINEAR DIOPHANTINE PROBLEM OF FROBENIUS. 1. Introduction

REFINED UPPER BOUNDS FOR THE LINEAR DIOPHANTINE PROBLEM OF FROBENIUS. 1. Introduction Version of 5/2/2003 To appear in Advanes in Applied Mathematis REFINED UPPER BOUNDS FOR THE LINEAR DIOPHANTINE PROBLEM OF FROBENIUS MATTHIAS BECK AND SHELEMYAHU ZACKS Abstrat We study the Frobenius problem:

More information

Cavity flow with surface tension past a flat plate

Cavity flow with surface tension past a flat plate Proeedings of the 7 th International Symposium on Cavitation CAV9 Paper No. ## August 7-, 9, Ann Arbor, Mihigan, USA Cavity flow with surfae tension past a flat plate Yuriy Savhenko Institute of Hydromehanis

More information

22.54 Neutron Interactions and Applications (Spring 2004) Chapter 6 (2/24/04) Energy Transfer Kernel F(E E')

22.54 Neutron Interactions and Applications (Spring 2004) Chapter 6 (2/24/04) Energy Transfer Kernel F(E E') 22.54 Neutron Interations and Appliations (Spring 2004) Chapter 6 (2/24/04) Energy Transfer Kernel F(E E') Referenes -- J. R. Lamarsh, Introdution to Nulear Reator Theory (Addison-Wesley, Reading, 1966),

More information

Differential Equations 8/24/2010

Differential Equations 8/24/2010 Differential Equations A Differential i Equation (DE) is an equation ontaining one or more derivatives of an unknown dependant d variable with respet to (wrt) one or more independent variables. Solution

More information

Concerning the Numbers 22p + 1, p Prime

Concerning the Numbers 22p + 1, p Prime Conerning the Numbers 22p + 1, p Prime By John Brillhart 1. Introdution. In a reent investigation [7] the problem of fatoring numbers of the form 22p + 1, p a, was enountered. Sine 22p + 1 = (2P - 2*

More information

Design and Development of Three Stages Mixed Sampling Plans for Variable Attribute Variable Quality Characteristics

Design and Development of Three Stages Mixed Sampling Plans for Variable Attribute Variable Quality Characteristics International Journal of Statistis and Systems ISSN 0973-2675 Volume 12, Number 4 (2017), pp. 763-772 Researh India Publiations http://www.ripubliation.om Design and Development of Three Stages Mixed Sampling

More information

Average Rate Speed Scaling

Average Rate Speed Scaling Average Rate Speed Saling Nikhil Bansal David P. Bunde Ho-Leung Chan Kirk Pruhs May 2, 2008 Abstrat Speed saling is a power management tehnique that involves dynamially hanging the speed of a proessor.

More information

Most results in this section are stated without proof.

Most results in this section are stated without proof. Leture 8 Level 4 v2 he Expliit formula. Most results in this setion are stated without proof. Reall that we have shown that ζ (s has only one pole, a simple one at s =. It has trivial zeros at the negative

More information

Ordered fields and the ultrafilter theorem

Ordered fields and the ultrafilter theorem F U N D A M E N T A MATHEMATICAE 59 (999) Ordered fields and the ultrafilter theorem by R. B e r r (Dortmund), F. D e l o n (Paris) and J. S h m i d (Dortmund) Abstrat. We prove that on the basis of ZF

More information

Case I: 2 users In case of 2 users, the probability of error for user 1 was earlier derived to be 2 A1

Case I: 2 users In case of 2 users, the probability of error for user 1 was earlier derived to be 2 A1 MUTLIUSER DETECTION (Letures 9 and 0) 6:33:546 Wireless Communiations Tehnologies Instrutor: Dr. Narayan Mandayam Summary By Shweta Shrivastava (shwetash@winlab.rutgers.edu) bstrat This artile ontinues

More information

Fiber Optic Cable Transmission Losses with Perturbation Effects

Fiber Optic Cable Transmission Losses with Perturbation Effects Fiber Opti Cable Transmission Losses with Perturbation Effets Kampanat Namngam 1*, Preeha Yupapin 2 and Pakkinee Chitsakul 1 1 Department of Mathematis and Computer Siene, Faulty of Siene, King Mongkut

More information

Estimating the probability law of the codelength as a function of the approximation error in image compression

Estimating the probability law of the codelength as a function of the approximation error in image compression Estimating the probability law of the odelength as a funtion of the approximation error in image ompression François Malgouyres Marh 7, 2007 Abstrat After some reolletions on ompression of images using

More information

Hankel Optimal Model Order Reduction 1

Hankel Optimal Model Order Reduction 1 Massahusetts Institute of Tehnology Department of Eletrial Engineering and Computer Siene 6.245: MULTIVARIABLE CONTROL SYSTEMS by A. Megretski Hankel Optimal Model Order Redution 1 This leture overs both

More information

Frequency hopping does not increase anti-jamming resilience of wireless channels

Frequency hopping does not increase anti-jamming resilience of wireless channels Frequeny hopping does not inrease anti-jamming resiliene of wireless hannels Moritz Wiese and Panos Papadimitratos Networed Systems Seurity Group KTH Royal Institute of Tehnology, Stoholm, Sweden {moritzw,

More information

arxiv:math/ v1 [math.ca] 27 Nov 2003

arxiv:math/ v1 [math.ca] 27 Nov 2003 arxiv:math/011510v1 [math.ca] 27 Nov 200 Counting Integral Lamé Equations by Means of Dessins d Enfants Sander Dahmen November 27, 200 Abstrat We obtain an expliit formula for the number of Lamé equations

More information

Modeling of discrete/continuous optimization problems: characterization and formulation of disjunctions and their relaxations

Modeling of discrete/continuous optimization problems: characterization and formulation of disjunctions and their relaxations Computers and Chemial Engineering (00) 4/448 www.elsevier.om/loate/omphemeng Modeling of disrete/ontinuous optimization problems: haraterization and formulation of disjuntions and their relaxations Aldo

More information

A Queueing Model for Call Blending in Call Centers

A Queueing Model for Call Blending in Call Centers A Queueing Model for Call Blending in Call Centers Sandjai Bhulai and Ger Koole Vrije Universiteit Amsterdam Faulty of Sienes De Boelelaan 1081a 1081 HV Amsterdam The Netherlands E-mail: {sbhulai, koole}@s.vu.nl

More information

Final Review. A Puzzle... Special Relativity. Direction of the Force. Moving at the Speed of Light

Final Review. A Puzzle... Special Relativity. Direction of the Force. Moving at the Speed of Light Final Review A Puzzle... Diretion of the Fore A point harge q is loated a fixed height h above an infinite horizontal onduting plane. Another point harge q is loated a height z (with z > h) above the plane.

More information

ON-LINE ESTIMATION OF THE VENTILATION RATE OF GREENHOUSES

ON-LINE ESTIMATION OF THE VENTILATION RATE OF GREENHOUSES ON-LINE ESTIMATION OF THE VENTILATION RATE OF GREENHOUSES J. Bontsema *, E.J. van Henten *, J.G. Kornet *, J. Budding ** and T. Rieswijk ** * Agrotehnology and Food Innovations B.V., Greenhouse Tehnology

More information

On the Complexity of the Weighted Fused Lasso

On the Complexity of the Weighted Fused Lasso ON THE COMPLEXITY OF THE WEIGHTED FUSED LASSO On the Compleity of the Weighted Fused Lasso José Bento jose.bento@b.edu Ralph Furmaniak rf@am.org Surjyendu Ray rays@b.edu Abstrat The solution path of the

More information

Singular Event Detection

Singular Event Detection Singular Event Detetion Rafael S. Garía Eletrial Engineering University of Puerto Rio at Mayagüez Rafael.Garia@ee.uprm.edu Faulty Mentor: S. Shankar Sastry Researh Supervisor: Jonathan Sprinkle Graduate

More information

INFLUENCE OF OPERATING AND CONSTRUCTION PARAMETERS ON THE BEHAVIOR OF HYDRAULIC CYLINDER SUBJECTED TO JERKY MOTION

INFLUENCE OF OPERATING AND CONSTRUCTION PARAMETERS ON THE BEHAVIOR OF HYDRAULIC CYLINDER SUBJECTED TO JERKY MOTION Proeedings of ICFDP 8: 8 th International Congress of Fluid Dynamis & Propulsion Deember 14-17, 006, Sharm El-Shiekh, Sinai, Egypt ICFDP8-EG-154 INFLUENCE OF OPERATING AND CONSTRUCTION PARAMETERS ON THE

More information

A Functional Representation of Fuzzy Preferences

A Functional Representation of Fuzzy Preferences Theoretial Eonomis Letters, 017, 7, 13- http://wwwsirporg/journal/tel ISSN Online: 16-086 ISSN Print: 16-078 A Funtional Representation of Fuzzy Preferenes Susheng Wang Department of Eonomis, Hong Kong

More information

Computer Science 786S - Statistical Methods in Natural Language Processing and Data Analysis Page 1

Computer Science 786S - Statistical Methods in Natural Language Processing and Data Analysis Page 1 Computer Siene 786S - Statistial Methods in Natural Language Proessing and Data Analysis Page 1 Hypothesis Testing A statistial hypothesis is a statement about the nature of the distribution of a random

More information

Scalable Positivity Preserving Model Reduction Using Linear Energy Functions

Scalable Positivity Preserving Model Reduction Using Linear Energy Functions Salable Positivity Preserving Model Redution Using Linear Energy Funtions Sootla, Aivar; Rantzer, Anders Published in: IEEE 51st Annual Conferene on Deision and Control (CDC), 2012 DOI: 10.1109/CDC.2012.6427032

More information

Arithmetic Circuits. Comp 120, Spring 05 2/10 Lecture. Today s BIG Picture Reading: Study Chapter 3. (Chapter 4 in old book)

Arithmetic Circuits. Comp 120, Spring 05 2/10 Lecture. Today s BIG Picture Reading: Study Chapter 3. (Chapter 4 in old book) omp 2, pring 5 2/ Leture page Arithmeti iruits Didn t I learn how to do addition in the seond grade? UN ourses aren t what they used to be... + Finally; time to build some serious funtional bloks We ll

More information

After the completion of this section the student should recall

After the completion of this section the student should recall Chapter I MTH FUNDMENTLS I. Sets, Numbers, Coordinates, Funtions ugust 30, 08 3 I. SETS, NUMERS, COORDINTES, FUNCTIONS Objetives: fter the ompletion of this setion the student should reall - the definition

More information

Intermediate Math Circles November 18, 2009 Solving Linear Diophantine Equations

Intermediate Math Circles November 18, 2009 Solving Linear Diophantine Equations 1 University of Waterloo Faulty of Mathematis Centre for Eduation in Mathematis and Computing Intermediate Math Cirles November 18, 2009 Solving Linear Diophantine Equations Diophantine equations are equations

More information

IMPEDANCE EFFECTS OF LEFT TURNERS FROM THE MAJOR STREET AT A TWSC INTERSECTION

IMPEDANCE EFFECTS OF LEFT TURNERS FROM THE MAJOR STREET AT A TWSC INTERSECTION 09-1289 Citation: Brilon, W. (2009): Impedane Effets of Left Turners from the Major Street at A TWSC Intersetion. Transportation Researh Reord Nr. 2130, pp. 2-8 IMPEDANCE EFFECTS OF LEFT TURNERS FROM THE

More information

3 Tidal systems modelling: ASMITA model

3 Tidal systems modelling: ASMITA model 3 Tidal systems modelling: ASMITA model 3.1 Introdution For many pratial appliations, simulation and predition of oastal behaviour (morphologial development of shorefae, beahes and dunes) at a ertain level

More information

Lecture 11 Buckling of Plates and Sections

Lecture 11 Buckling of Plates and Sections Leture Bukling of lates and Setions rolem -: A simpl-supported retangular plate is sujeted to a uniaxial ompressive load N, as shown in the sketh elow. a 6 N N a) Calulate and ompare ukling oeffiients

More information

Aircraft CAS Design with Input Saturation Using Dynamic Model Inversion

Aircraft CAS Design with Input Saturation Using Dynamic Model Inversion International Journal of Control, Automation, and Systems Vol., No. 3, September 003 35 Airraft CAS Design with Input Saturation sing Dynami Model Inversion Sangsoo Lim and Byoung Soo Kim Abstrat: This

More information

On the Relative Value of Local Scheduling versus Routing in Parallel Server Systems

On the Relative Value of Local Scheduling versus Routing in Parallel Server Systems On the Relative Value of Loal Sheduling versus Routing in Parallel Server Systems Rong Wu as.om 343 Thornall Street Edison, NJ, USA 08837 Douglas G. Down Department of Computing and Software MMaster University

More information

ONLINE APPENDICES for Cost-Effective Quality Assurance in Crowd Labeling

ONLINE APPENDICES for Cost-Effective Quality Assurance in Crowd Labeling ONLINE APPENDICES for Cost-Effetive Quality Assurane in Crowd Labeling Jing Wang Shool of Business and Management Hong Kong University of Siene and Tehnology Clear Water Bay Kowloon Hong Kong jwang@usthk

More information

New Strategy for Doubling-Free Short Addition-Subtraction Chain

New Strategy for Doubling-Free Short Addition-Subtraction Chain Applied Mathematics & Information Sciences 2(2) (2008), 123 133 An International Journal c 2008 Dixie W Publishing Corporation, U. S. A. New Strategy for Doubling-Free Short Addition-Subtraction Chain

More information

Methods of evaluating tests

Methods of evaluating tests Methods of evaluating tests Let X,, 1 Xn be i.i.d. Bernoulli( p ). Then 5 j= 1 j ( 5, ) T = X Binomial p. We test 1 H : p vs. 1 1 H : p>. We saw that a LRT is 1 if t k* φ ( x ) =. otherwise (t is the observed

More information

Wave Propagation through Random Media

Wave Propagation through Random Media Chapter 3. Wave Propagation through Random Media 3. Charateristis of Wave Behavior Sound propagation through random media is the entral part of this investigation. This hapter presents a frame of referene

More information

On Certain Singular Integral Equations Arising in the Analysis of Wellbore Recharge in Anisotropic Formations

On Certain Singular Integral Equations Arising in the Analysis of Wellbore Recharge in Anisotropic Formations On Certain Singular Integral Equations Arising in the Analysis of Wellbore Reharge in Anisotropi Formations C. Atkinson a, E. Sarris b, E. Gravanis b, P. Papanastasiou a Department of Mathematis, Imperial

More information

International Journal of Electronics and Computer Science Engineering 817. Available Online at ISSN

International Journal of Electronics and Computer Science Engineering 817. Available Online at   ISSN International Journal of Eletronis and Computer Siene Engineering 817 Available Online at www.ijese.org ISSN- 2277-1956 A Duly Synhronized, Straightforward Approah For Realizing the General Charateristis

More information

The simulation analysis of the bridge rectifier continuous operation in AC circuit

The simulation analysis of the bridge rectifier continuous operation in AC circuit Computer Appliations in Eletrial Engineering Vol. 4 6 DOI 8/j.8-448.6. The simulation analysis of the bridge retifier ontinuous operation in AC iruit Mirosław Wiślik, Paweł Strząbała Kiele University of

More information

Development of Fuzzy Extreme Value Theory. Populations

Development of Fuzzy Extreme Value Theory. Populations Applied Mathematial Sienes, Vol. 6, 0, no. 7, 58 5834 Development of Fuzzy Extreme Value Theory Control Charts Using α -uts for Sewed Populations Rungsarit Intaramo Department of Mathematis, Faulty of

More information

Chapter 8 Hypothesis Testing

Chapter 8 Hypothesis Testing Leture 5 for BST 63: Statistial Theory II Kui Zhang, Spring Chapter 8 Hypothesis Testing Setion 8 Introdution Definition 8 A hypothesis is a statement about a population parameter Definition 8 The two

More information

Sensor management for PRF selection in the track-before-detect context

Sensor management for PRF selection in the track-before-detect context Sensor management for PRF seletion in the tra-before-detet ontext Fotios Katsilieris, Yvo Boers, and Hans Driessen Thales Nederland B.V. Haasbergerstraat 49, 7554 PA Hengelo, the Netherlands Email: {Fotios.Katsilieris,

More information

Advanced Computational Fluid Dynamics AA215A Lecture 4

Advanced Computational Fluid Dynamics AA215A Lecture 4 Advaned Computational Fluid Dynamis AA5A Leture 4 Antony Jameson Winter Quarter,, Stanford, CA Abstrat Leture 4 overs analysis of the equations of gas dynamis Contents Analysis of the equations of gas

More information

5.1 Composite Functions

5.1 Composite Functions SECTION. Composite Funtions 7. Composite Funtions PREPARING FOR THIS SECTION Before getting started, review the following: Find the Value of a Funtion (Setion., pp. 9 ) Domain of a Funtion (Setion., pp.

More information

MATHEMATICAL AND NUMERICAL BASIS OF BINARY ALLOY SOLIDIFICATION MODELS WITH SUBSTITUTE THERMAL CAPACITY. PART II

MATHEMATICAL AND NUMERICAL BASIS OF BINARY ALLOY SOLIDIFICATION MODELS WITH SUBSTITUTE THERMAL CAPACITY. PART II Journal of Applied Mathematis and Computational Mehanis 2014, 13(2), 141-147 MATHEMATICA AND NUMERICA BAI OF BINARY AOY OIDIFICATION MODE WITH UBTITUTE THERMA CAPACITY. PART II Ewa Węgrzyn-krzypzak 1,

More information