M.S. Bazaraa, J.J. Jarvis, and H.D. Sherali (1990). Linear Programming and Network Flows, Wiley, New York.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....we proved that it is NP hard and NP complete on general graphs and planar graphs respectively. We proposed a heuristic algorithm with time complexity O(jV j log jV j) to solve the vertex compression problem. And we also proposed another heuristic algorithm by applying linear programming [2]. In order to get the optimal solution, we restrict the linear programming to the 0 1 integer programming [1] We also show some experiment results. In the future, we may develop a new algorithm with the branch and bound strategy [4] the simulated annealing strategy [6] or the genetic strategy ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Network Flows. John Wiley & Sons, 1990. 21

....function moving up through the region. The highest position of the cost line is obtained at vertex A. The theory of Linear Programming shows, that for a convex and bounded region, the objective function is maximized either at one of the vertices, or on a line segment joining two vertices Bazaara [1]. Thus, we can always pick one or more best performing classifiers. Figure 6. Cost functions traversing the ROC convex hull As in the ROCCH analysis, we have iso performing lines, which help visualize performance of classifiers under various cost structures. In the Linear Programming setting, ....

....Maximize C = c x j Subject to a ij x b i i = 1, k. x j 0 j = 1, d The theory of linear programming assures us, that if the set of constraints forms a convex and bounded set (called the feasible region) then an optimal solution is found on the boundaries of the feasible region [1]. In our two dimensional case, a solution can be found at a vertex, or on an edge joining two vertices. Note that the feasible set is created as a conjunction of several linear inequalities. Not all vertices are known explicitly. A number of computational techniques have been designed to find ....

[Article contains additional citation context not shown here]

Bazaraa, Moktar S. at al. Linear Programming and Network Flows. John Wiley & Sons, Inc., 1997.

....it will serve as a good benchmark to compare our heuristics against. 4 Formalization as a multi commodity flows problem We also attempted to map Appia s problem onto the well known theoretical problem of multicommodity flows. An introduction to the multi commodity flows problem can be culled from (Bazaraa, Jarvis and Sherali, 1990). Boesch, 1976) also gives a good survey to the wide array of existing graph problems, and a few interesting papers on multicommodity flows. k ij k ji 0 for internal nodes , k , for terminal nodes = flow conservation constraint) f k ji h h i cp s s f k ij ij x (node ....

Mokhtar S. Bazaraa, John J. Jarvis and Hanif D. Sherali, "Linear programming and Network flows", Chapter 12, John Wiley and Sons, 1990.

....with both bb and una occurs when an integer solution is sought and the constraints represent an unbounded polyhedron with no integer solutions. The two algorithms will never terminate with the empty solution. Informal experiments, however, show that a cutting plane method such as Gomory cuts [1] in combination with branchand bound seems to solve this problem. 8. CONCLUSIONS una is highly dependent on the composition of the input. We have seen that it is very easy to construct systems that makes the iteration count increase drastically: system (10) requires more then 300 times the ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali. Linear Programming and Network Flows. John Wiley & Sons, 2nd edition, 1990.

....area in the field of mathematical optimization, where a linear function z = c 1 x 1 c 2 x 2 Delta Delta Delta c n x n is optimized, subject to a set of linear equality or inequality constraints. The Simplex is the most widely used solution method for linear programming problems [3, 18, 21]. Since its introduction by George B. Dantzig in 1947 [7] the Simplex has been extensively used to solve a wide base of optimization and resource allocation problems in the government, military, and the industry. As in the solution of any large scale mathematical system, the computation time for ....

Bazaraa, M. S., J. J. Jarvis and H. D. Sherali, Linear Programming and Network Flows, John Wiley & Sons, Inc., 1990.

....to a dependence vector of the iteration space. The corner points of this convex hull form the set of extreme points for the convex solution space. These extreme points have the property that any point in the convex space can be represented as a convex combination of these extreme points [10]. The dependence vectors of these extreme points form a set of extreme vectors for the dependence vector set [11] We compute the minimum dependence distances from these extreme vectors. Using these minimum dependence distances we tile the iteration space. For parallel execution of these tiles, ....

....and theorems through which we can find the minimum and maximum dependence distances. We use a theorem from linear programming that states For any linear function which is valid over a bounded and closed convex space, its maximum and minimum values occur at the extreme points of the convex space [10, 17]. Theorem 1 is based on the above principle. Since both d i (x; y) and d j (x; y) are linear functions and are valid over the IDCH, we use this theorem to compute the minimum and maximum dependence distances in both i and j dimensions. Theorem 3 : The minimum and maximum values of the dependence ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Network Flows. John Wiley & sons, 1990.

....a linear combination of variables, given a set of constraints. Linear programming is used extensively in economics and engineering. Examples from economics include production scheduling and the determination of shadow prices i.e. the price a manufacturer would pay for a given scarce resource ([4]) An example of an engineering application would be profit maximization in a factory that manufactures a number of di#erent products from the same raw material using the same resources. Linear programming can be carried out using the simplex method described by Wood and Dantzig in 1949 ( 9] The ....

M. S. Bazaraa, J. J. Jarvis and H. D. Sherali, "Linear Programming and network flows", John Wiley & sons, 1977.

....approach. First, the link assignment problem is formulated as a mixed integer linear programming problem whose solution gives the globally optimal link assignment. Second, an iterative optimization method is used to solve this mixed integer linear problem, which is known to be an NP hard problem [7]. The iterative method is based on the simulated annealing [8] that is commonly used in solving many other combinatorial optimization problems. It refines the link assignment incrementally and yields a near optimal link assignment. The remainder of this paper is organized as follows: The next ....

....The link assignment problem formulated in this way is a mixed integer linear programming problem since the formulation has both integer and real variables. This problem is known to be NP hard, and thus, its time complexity increases exponentially as the number of variables or constraints increases [7]. Start Make a new link assignment (by applying branch and exchange) Compute the maximum link utilizatio (by solving the static routing problem) If refinement possible Branch and Exchange End Fig. 4. Flow Diagram of an Iterative Optimization Approach B. An Iterative Optimization ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Networks Flows. Toronto: John Wiley & Sons, 1990.

....partition P = P0 U P U . U P, induced by the initial macromarking is disjoint, i.e. P f Pj, 0 for all j j . Under assumption A4, it is easy to prove that V is unimodular, thus an integer solution of the integer problem above can also be found by the computationally efficient simplex method [1]. Furthermore, the columns of V are orthogonal, and a closed form solution of this problem can be given. Proposition 5 ( 7] Consider the integer problem max ff M s.t. V T . M : V T . 3w B Here V, 13, and 3)I are given as it, Algorithm 3 (thus we assume the problem is feasible) P : Po U P ....

....The initial estimate computed by Algorithm 3 is Ms0 = 0 0 0] T, B 0 = 3] thus the initial marking T T T of he oserver ne (shown in Figure 1. is [ o 0 se consisen markings is (w0 ] M ] M(p) M(p) M(ps) 3) Ift fires from M0, he oserved word is w = t and he system reaches he markin M = [2 1 0] T, T T T B] 1 2] T. The of w consisen while he oserver ne reaches he new markin [ 0 0 markings is (w ] M ] M(p) M(p) M(ps) 3, M(p) 1) 4 Control using observers In this section, we show how the marking estimate computed by an observer can be used by a control agent to enforce a ....

[Article contains additional citation context not shown here]

M.S. Bazaraa, J.J. Jarvis, Linear Programming and Network Flows, John Wiley & Sons, 1977.

....to a dependence vector of the iteration space. The corner points of this convex hull form the set of extreme points for the convex solution space. These extreme points have the property that any point in the convex space can be represented as a convex combination of these extreme points [8]. The dependence vectors of these extreme points form a set of extreme vectors for the dependence vector set [9] We compute the minimum dependence distances from these extreme vectors. Using these minimum dependence distances we tile the iteration space. For parallel execution of these tiles, ....

....and theorems through which we can find the minimum and maximum dependence distances. We use a theorem from linear programming that states For any linear function which is valid over a bounded and closed convex space, its maximum and minimum values occur at the extreme points of the convex space [8, 17]. Theorem 1 is based on the above principle. Since both d i (x; y) and d j (x; y) are linear functions and are valid over the IDCH, we use this theorem to compute the minimum and maximum dependence distances in both i and j dimensions. Theorem 1 : The minimum and maximum values of the dependence ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Network Flows. John Wiley & sons, 1990.

....chosen items. This is an NP hard problem, but can be solved relatively efficiently, both in theory and practice. From a practical point of view, several good heuristics have been discovered for MCKP. One class of heuristics is based on solving the underlying linear programming (LP) relaxation [2] [3], 10] 20] 30] and rounding the (possibly) fractional solution obtained to an integral one. Such rounding heuristics construct solutions with values that are provably within a factor of 2 of the best possible; perhaps more importantly, such heuristics perform extremely well in practice [2] ....

....the total number of fractional variables is 12 22.ff The number of integral variables for both (C1) and (C2) is 12 2nff . Further, the number of non negative variables in an optimal basic feasible solution (BFS) of a linear program is bounded above by the number of constraints in the problem [3]. Hence we get, n 1212 i=11 n 12 i=11 21( #(fractional integral vars. 22) 2) Combining, 1( n ik k n ik k nnnffnff ffnn = A similar counting argument on F2 gives that 12 1 1 n k k ffn = Therefore, although F1 bounds the feasible integral ....

M. Bazaraa, J. J. Jarvis, and H. D. Sherali, "Linear Programming and Network Flows", 2 nd Edition, John Wiley and Sons, 1990.

....problem, there are some disjoint sets of objects. Objects in each of these sets have values and weights associated with them, and exactly one object from each set must be chosen without violating the knapsack size. This is known to be an NP hard problem, but its linear programming (LP) relaxation [3], 4] 10] can be solved extremely quickly (in linear time) The linear time algorithm [1] 6] 18] is based on using dual descent in the LP dual problem. Another approach for solving the linear relaxed problem is by identifying the convex hull of operating points in each set [8] and this has ....

M. Bazaraa, J. J. Jarvis, and H. D. Sherali, "Linear Programming and Network Flows", 2 nd Edition, John Wiley and Sons, 1990.

.... any 2 Theta 2 matrix A, then det(A) vec(A) Deltavec(A) Phi = O O Matrix of the moments of the centered observations = Theta ff 1 Delta Delta Delta ff p 1 Omega I [2] Matrix of the camera scale factors Omega ijkl A 4 way array Omega A vector such that Omega q [4] = Omega ijkl q i q j q k q l , using Einstein convention Pi k n A fixed n k Theta n k permutation matrix such that, for any n Theta n matrix A and any integer k then vec i A [k] j = Pi k n vec(A) k] Xi = C h I [p 1 ] Gamma CX(CX) i C Auxiliary matrix, depending ....

....a concave sub stochastic equivalent that can be found using the techniques in Appendix G. The demonstration builds the polynomial function explicitly. Theorem 5 The problem 4.21 is equivalent to P = arg min P J 6 (X; P; Y) s.t. P 2 P p (p 1 ; p 2 ) 4. 23) with J 6 (X; P; Y) Omega q [4] (4:24) where Omega is independent of Q and q [4] q Omega q Omega q Omega q. 20 Proof: We start by using the result A.7 with M = X and N = Y then L = O, so arg min Q J 5 (X; Q; Y) arg min Q det( Phi) arg min Q det i O O j = arg min Q h det i X C CX j ....

[Article contains additional citation context not shown here]

M. Bazaraa, J. Jarvis, and H. Sherali. Linear Programming and Network Flows. Wiley, second edition, 1990.

....inequality (17) becomes an equality. then Problem 3 achieves an optimal solution jE 1 j k=1 f(ek ) dk . Proof. Obvious from the above discussion. 3.5 Solving Problem 3 Here, let us consider each vector w(v i ) q i and dk as a single computation unit. Based on the duality theory [24, 2], Problem 3, excluding the constraint (9) is equivalent to max jE 1 j k=1 ( f(ek ) dk ) 18) subject to e k = v i 2E 1 f(ek ) w(v i ) e k = v i ; 2E 1 f(ek ) 1 i jV1 j: 19) f(e i ) 0; 1 i jE1 j: 20) The constraint (9) is mandatory for the equivalence between ....

....equal, i.e. jV 1 j i=1 w(v i ) q i = jE 1 j k=1 f(ek ) dk holds. If we can prove that the constraint (9) holds for the optimal solution of Problem 4, such a solution must also be optimal for Problem 3, according to Theorem 4. There exist plenty of algorithms to solve Problem 4 [1, 2]. Although those algorithms are targeted to the scalar system (the vector length equals to 1) some of them can be directly adapted to our system by vector summation, subtraction and comparison operations. A network simplex algorithm [2] can be directly utilized to solve our system. The ....

[Article contains additional citation context not shown here]

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali. Linear Programming and Network Flows. Wiley, New York, 1990.

....for the value of C:lower. These literals correspond to the set cl in the bound con ict clause. Applying this reasoning to both assignments of a given variable, allows implementing non chronological backtracking and cl can be build as: cl = fl : l = 0 l 2 i i 2 Sg (11) 2 See [8] for a de nition of slack and arti cial variables. 8 5.2 Reducing Dependencies in LPR Bound Con icts The previous section describes how to obtain an explanation on the value of C:lower when using LPR. However, a more careful analysis allows the identi cation of some situations where literals ....

J. J. J. M. S. Bazaraa and H. D. Sherali. Linear Programming and Network Flows. 2nd Ed., John Wiley & Sons, 1989.

....anticipated demands. In all such examples, the flow of supplies to satisfy demands is carried out subject to certain capacity limitations. Deterministic network flow models have numerous additional applications, and a variety of efficient solution algorithms have been developed for this problem [4, 26]. More realistically, the elements of such a flow network should be viewed as stochastic, rather than deterministic, since the supplies, demands, and transmission capacities are rarely known with certainty. A stochastic version of the standard network flow problem is the focus of this paper. One ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Network Flows. John Wiley, New York (1990).

....anticipated demands. In all such examples, the flow of supplies to satisfy demands is carried out subject to certain capacity limitations. Deterministic network flow models have numerous additional applications, and a variety of efficient solution algorithms have been developed for this problem [4, 26]. More realistically, the elements of such a flow network should be viewed as stochastic, rather than deterministic, since the supplies, demands, and transmission capacities are rarely known with certainty. A stochastic version of the standard network flow problem is the focus of this paper. One ....

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Network Flows. John Wiley, New York (1990).

No context found.

M.S. Bazaraa, J.J. Jarvis, and H.D. Sherali (1990). Linear Programming and Network Flows, Wiley, New York.

No context found.

M. S. Bazaraa, J. J. Jarvis and H. D. Sherali, Linear Programming and Network Flows, 2nd ed., John Wiley & Sons, New York, 1990. 9

No context found.

M.S. Bazara, J.J. Jarvis, H.D. Sherali, "Linear Programming and Network Flows", 1990, John Wiley & Sons, ISBN 0-471-63681-9

No context found.

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali, Linear Programming and Network Flows. John Wiley & Sons, January 1990.

No context found.

M. S. Bazaraa, J. J. Jarvis, and H. D. Sherali. Linear Programming and Network Flows. John Wiley & sons, 1990.

No context found.

M.S. BAZARAA, J.J. JARVIS and H.D. SHERALI, Linear Programming and Network Flows, 2 nd ed., Willey, New York, 1990.

No context found.

Mokhtar S. Bazaraa, John J. Jarvis, and Hanif D. Sherali. Linear Programming and Network Flows. John Wiley and Sons, New York, 2nd edition, 1990.

No context found.

M. Bazaraa, J. Jarvis, H. Sherali, Linear Programming and Network Flows, 2nd Edition, John Wiley and Sons, 1990.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC