LOCAL SEARCH ALGORITHMS FOR THE MAXIMUM K-PLEX PROBLEM

Size: px
Start display at page:

Download "LOCAL SEARCH ALGORITHMS FOR THE MAXIMUM K-PLEX PROBLEM"

Transcription

1 LOCAL SEARCH ALGORITHMS FOR THE MAXIMUM K-PLEX PROBLEM By ERIKA J. SHORT A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA

2 c 2008 Erika J. Short 2

3 To my parents, Tim and Jeanne, and my sisters, Alicia and Kayla. 3

4 ACKNOWLEDGMENTS I would like to thank my advisor, Dr. Panos M. Pardalos, for all of his encouragement and support. Also, I would like to acknowledge Dr. J. Cole Smith for serving on my committee and for all of his honest advice. Lastly, I would like to thank my family and friends for their never-ending moral support. 4

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS LIST OF TABLES LIST OF FIGURES ABSTRACT CHAPTER 1 INTRODUCTION THE MAXIMUM K-PLEX PROBLEM Foundations The Maximum k-plex Problem HEURISTIC APPROACH Development Heuristics Greedy Algorithm Tabu Search Method Computational results Exchange Algorithm Method Computational results CONCLUSIONS REFERENCES BIOGRAPHICAL SKETCH

6 Table LIST OF TABLES page 3-1 The tested DIMACS instances Tabu results for k = 1 and k = Exchange results for k = 1 and k = Result comparison for k = Result comparison for k =

7 Figure LIST OF FIGURES page 3-1 Greedy algorithm for k-plex Tabu algorithm for k-plex Exchange algorithm for k-plex

8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science LOCAL SEARCH ALGORITHMS FOR THE MAXIMUM K-PLEX PROBLEM Chair: Panagote M. Pardalos Major: Industrial and Systems Engineering By Erika J. Short May 2008 The maximum k-plex problem is one of many ways to classify a cohesive subgroup in social network analysis. Considered a type of relaxed clique, the maximum k-plex problem is a degree based approach to identifying closely related vertices in a graph. The recent development of applying optimization to the maximum k-plex problem has the capability for growth. Local search algorithms were developed and applied to benchmark problems for the maximum k-plex problem. Computational experiments were performed, and new lower bounds have been successfully attained. 8

9 CHAPTER 1 INTRODUCTION The application of graph theory to the notion of social networks has been developed for many years. Unsurprisingly, the effects of globalization along with the an increase of interest in terrorism have propelled a new surge of interest in social network theory. Along with the study of social networks, certain mathematical models have been revived from the past. One such example is the maximum k-plex problem. The maximum k-plex problem has roots in the study of cohesive subgroups of a network. According to [25], the concept of cohesive subgroups comes from the close relationships between actors in a social network. By studying these groups, sociologists can form theories that explain the behavior of actors in social networks. The most famous of these cohesive subgroup models is known as the maximum clique problem. Although it is well-known, so are its difficulties. The restrictions associated with maximum clique have propelled the study and formation of clique relaxations. These relaxed approaches give way to new opportunities by focusing on specific properties of cohesive subgroups. In particular, the maximum k-plex problem focuses on the reachability and robustness of the graph, and relaxes the property of familiarity, as stated in [3]. This thesis will apply heuristic techniques to the maximum k-plex problem. Chapter 2 will cover the background of the problem. In chapter 3, heuristic approaches and results are addressed. Chapter 4 discusses conclusions and the future development of techniques that can be applied to the maximum k-plex problem. 9

10 CHAPTER 2 THE MAXIMUM K-PLEX PROBLEM 2.1 Foundations The maximum clique problem is well-known to be NP -hard. A clique can be represented on an undirected graph as a subset of nodes that are all pairwise adjacent. Attaining the maximum clique requires finding the largest clique that exists in the graph. With applications in many fields, the maximum clique problem has been extensively researched in the past [21]. Many sources have pointed out problems with the maximum clique problem, regarding its restrictiveness [25], modelling disadvantages [23] [10], and difficult approximation [5]. In light of these problems, researchers developed many different relaxations to the maximum clique problem. Since cliques can be described as cohesive subgroups of a social network, certain properties can be used to describe their nature: familiarity (each vertex has many neighbors), reachability (a low diameter), and robustness (high connectivity) [3]. Each of these properties has its own difficulties, resulting in different relaxation techniques. Some well-known relaxation models are k-clique [18], k-club [2], and k-clan [19]. This paper considers another model called the maximum k-plex problem. This particular relaxation is degree oriented and was developed by Seidman and Foster in 1978 [23]. 2.2 The Maximum k-plex Problem Let G = (V, E) be a simple, undirected graph, where V = {1, 2,..., n}, E = (i, j) such that i, j V, and the degree, deg G (v i ), is number of neighbors of v i in G. Given a subset, S V, let G[S] be the induced subgraph of S. If Equation 2 1 holds, the set S induces a k-plex. deg G[S] (v i ) S k v i S (2 1) 10

11 Although the maximum k-plex problem was developed years ago, studies in optimization have neglected it until recently. In [3], the problem was reintroduced, considering its computational complexity and developing an integer programming formulation. In regards to complexity, when k is arbitrary, the maximum clique problem exists as a special case of maximum k-plex and is clearly NP -hard [3]. In addition, an NP -completeness proof for finding a k-plex of a fixed value of k in a graph was published [3]. Also, in [3] the first known computational results for the maximum k-plex problem was presented. Starting with an integer programming formulation, they constructed a branch-and-cut method to be applied to well-known data sets. A peeling procedure used to remove vertices with low degree was also used in preprocessing. Their benchmark solutions were obtained by implementing the method using ILOG CPLEX 9.0R[ILOG] software. No other published results are known to exist. 11

12 CHAPTER 3 HEURISTIC APPROACH 3.1 Development In optimization, the maximum k-plex problem has few computational results. A branch-and-cut algorithm that produced benchmark results for the problem was published in [3]. Branch-and-cut is considered an exact algorithm that uses cutting plane methods to increase the performance of a branch-and-bound algorithm. Although this method is exact, its performance is very slow. Therefore, approximation methods can be used to find solutions at a much faster rate. Since the maximum k-plex problem is relatively unexplored, heuristic methods have been developed to produce new benchmark solutions. Although the branch-and-cut algorithm was tested on several different instances, the experiments in this study are tested only against DIMACS instances [9]. These can be found in Table Greedy Algorithm 3.2 Heuristics The initial approach to approximating the maximum k-plex begins with a simple greedy algorithm for the maximum clique problem. This method involves selecting the vertex with the highest degree and adding it to the current solution set. Iteratively, neighboring vertices are added to set if they have the highest degree of all nodes that can be added to the clique. Vertices are continually added until no more can be added to the set. With little effort, this approach can be expanded to the maximum k-plex problem. This adaptation requires a slightly more detailed validity consideration. Instead of requiring all vertices to be connected to the candidate vertex, the new solutions must ensure that they satisfy the k-plex requirements. By definition, all vertices in the current set, G[S], and the candidate vertex, v i, have to satisfy Equation 2 1 to retain k-plex feasibility. In order to implement the algorithm, each vertex is classified when it is added to the current solution.vertices are classified as 12

13 one of two groups: those with positive potential and those with none. Potential, denoted pot(v i ), can be described as the amount of flexibility that the vertex has within the solution set. Potential is calculated by taking the difference between the k-plex degree requirement and the number of neighbors that the candidate vertex has within the set, see Equation 3 1. This property is useful in quickly determining the quality of candidate solutions. Similar to maximum clique, all vertices with zero potential must be neighbors to any vertex selected to be added to the k-plex in order to retain feasibility. pot(v i ) = deg G[S] (v i ) S + k v i S (3 1) Also, this approach requires the chosen vertex to have the maximum degree amongst all candidates outside of the set, deg G[ S] (v i ). In addition, if tie-breaking is necessary, the added vertex should be selected at random. The implementation of this algorithm is shown in Figure 3-1. Basic greedy algorithms are known to produce good results, but they are usually not significant enough to report. Alone, the greedy algorithm has a tendency to include vertices of high degree, although those vertices may not belong to the maximum k-plex. For more dependable results, the use of more advanced search techniques are needed. Therefore, this greedy approach simply provides a platform to continue development of a more in-depth heuristic Tabu Search Method In order to improve results, local search techniques were considered. Local search methods are characterized by searching the neighborhood of a current solution to find improvements. Local search heuristics have been known to produce good results when applied to difficult optimization problems. Within the classification of local search methods, tabu search is considered to be the most effective. Developed by Fred Glover, tabu search is a method that limits the number 13

14 of future solutions based on a history of previous moves [12]. Unlike many other local search techniques, this method provides a way to escape local optima. With tabu search, a set of previously chosen moves are stored in a tabu list that disallows that move to be made again until a specified number of new moves have been taken. This limiting factor reduces the chance that the algorithm will continue to revisit the same local solution. Considering its past performance, tabu search was chosen as starting point for applying local search to the maximum k-plex problem. In order to implement tabu search, the qualities of the neighborhood structures need to be determined for efficient results [13]. For this method, the solutions that have been previously chosen using the greedy algorithm will be classified as tabu. These choices will be forbidden to be used again until a specified number of moves have been made. In such a greedy algorithm, the tabu solution will be overlooked, and the next best choice will be taken. The length of time that a move is classified as tabu is called its tabu tenure. Although there are many different ways of determining and adjusting the tabu tenure, most choices are based on empirical tests that begin with an initial set value for all problems [13]. For the results presented in this paper, the tabu tenure was chosen as a set value, with no adjustment within the algorithm itself. To implement the neighborhood search, each locally maximal solution needs to be changed. For this procedure, a local maximum is adjusted by removing a vertex from its current solution and adding new vertices based on the greedy heuristic. The selection of a deleted vertex is based on a partially random approach. Referencing to the values of potential for each vertex in the graph, vertices with zero potential are critical to the structure of the graph. Because these vertices are more restrictive to the solution, a random vertex from this subset is selected to be removed from the set. In such a case that a vertex cannot be selected with a potential of zero, the deleted vertex must be chosen from the rest of the set. This algorithm runs in a recursive manner, continuing until a stopping criteria is met. The pseudo code for the algorithm can be found in Figure

15 Computational results The tabu search method developed in this project was implemented using C programming language. Results were obtained using an AMD Opteron 2.2 GHz processor with 4 GB of RAM. Search parameters were equivalent for all experiments: tabu tenure = 10 iterations and run time = 1000 seconds. Although brief experiments were run for tabu tenure > 10, outputs provided insignificant results. The algorithm was applied to a variety of graphs from the benchmark clique instances of the Second DIMACS challenge [9]. Experiments on the graphs were done with values k = 1 (clique) and k = 2. Each experiment was performed for 20 independent runs. The results of the tests can be seen in Table 3-2, showing the average time to solution, maximum best solution, minimum best solution, and average best solution. These results are compared with later results and the benchmark solutions of [3] in Table 3-4 and Table 3-5. Comparison values indicate average time to solution and maximum best values. Values in square brackets are the bounds for non-optimal terminations in the benchmark instances from [3]. Values in parenthesis are new bounds. Experiments that did not obtain optimum (or current lower bound) are marked with an asterisk (*). The tabu search algorithm performed well on many problems, and new bounds were obtained for certain instances. Although 1-plexes are cliques, the algorithm provided improved results compared with previous k-plex results. According to DIMACS [9], the 1-plex bound attained for MANN 27.clq is the published size for maximum clique. Additionally, the algorithm found a new lower bound for hamming10-4.clq with k = 2. Although most results were matched or improved, some benchmark values were not attained Exchange Algorithm Method The proposed tabu heuristic uses a process of randomly deleting vertices and reevaluating the current solution set. New vertices are considered to be added to the 15

16 set if they satisfy the k-plex feasibility conditions for the smaller set and do not belong to the tabu list. Once a local maximum is attained, the deleted vertex is chosen at random. Only after a vertex is deleted, can new vertices be considered for addition to the k-plex with the greedy approach. It has been considered that deleting a random node can have varying affects on the future growth of the k-plex. Once the algorithm reaches a locally optimal solution, it must reduce the k-plex size to move within the neighborhood. Because the vertex is deleted randomly, the solution may not reach its previous size if no other vertices can replace the deleted vertex. Therefore, a local search method is considered that will prohibit moves within the neighborhood that reduce the size of the k-plex. In this method, local search is conducted through exchanging vertices. When a local maximum is attained, a vertex within the set will be replaced by an exchange partner outside of the solution set. The exchange partner is a vertex chosen randomly and evaluated for possible exchange moves within the set. In order to decrease the complexity of the algorithm, the first move that satisfies conditions will be selected. Once feasibility is established, the two vertices can be exchanged. The algorithm then continues in a greedy fashion, adding new vertices to the set. Finding an exchange pair consists of evaluating simple conditions. After a vertex outside of the solution is chosen, it is randomly compared against the vertices within the set. When choosing an exchange move, two considerations need to be made: k-plex feasibility of the chosen vertex and stability of the current set. The first condition is for the chosen vertex to have k-plex feasibility after being added to the set. From the basic k-plex condition in Equation 2 1, the exchange conditions can be extended in terms of the relationship between the two chosen vertices. Because the feasibility of the new vertex is being evaluated before removing its exchange partner, the k-plex conditions are dependent on the relationship between the two. If the vertices are neighbors, the evaluated conditions must be offset by one, in order to compensate for the degree change when the current vertex is removed from the set. This situation occurs 16

17 because the degree of the new vertex will consider its exchange partner as still belonging to the set. The k-plex conditions for evaluating an exchange when the two vertices are not neighbors can be found in Equation 3 2. The conditions for when they are neighbors can be found in Equation 3 3, where S is the current solution set, and v new is the vertex being considered for addition to the set. N(v new ) S S k (3 2) N(v new ) S S k + 1 (3 3) Once the vertex is known to satisfy k-plex conditions, it must be compared with the current vertex to ensure that it does not break the feasibility conditions for the rest of the set. Similar to the requirements of the greedy algorithm, the potential values of the current solution are used to ensure the structure of the k-plex. Again, critical vertices are those with potential equal to zero, and any new vertex being added to the set must be connected to these vertices. With the conditions of exchange, potential values can only be increased or remain the same when a vertex is removed from the set. Therefore, only vertices with zero potential that are also neighbors to the vertex being considered for removal need to be checked for set stability. In order to prevent the neighboring vertices from attaining negative potential (breaking the k-plex), they all must also be neighbors of the exchange partner. This condition can be seen in Equation 3 4, where S is the current solution set, S BL is the subset S with potential equal to zero, v new is the vertex being added to the set, and v old is the vertex being removed from the set. N(v new ) S BL N(v old ) S BL (3 4) After all conditions are satisfied, the exchange can be made. A new solution set is formed, allowing the greedy algorithm to continue checking for new solutions. The 17

18 process continues to repeat until a stopping criteria is specified. The pseudo code for this algorithm can be found in Figure 3-3. In addition to this exchange algorithm, another method, using tabu search, was considered. This method incorporated the tabu search heuristic into the selection of exchange pairs. The tabu method would prevent exchanges with vertices that had been exchanged recently. Preliminary experiments were performed with tabu tenures of various sizes. After these tests, results showed a significant increase in process time for the exchange algorithm with tabu, with either no increase or a decrease in solution size. Therefore, no further experiments were performed on this combined algorithm Computational results The exchange local search algorithm developed in this project was implemented using C programming language. Results were obtained using an AMD Opteron 2.2 GHz processor with 4 GB of RAM. All experiments had the stopping criteria of run time = 1000 seconds. The algorithm was applied to the same set of benchmark clique instances as the tab algorithm, from the Second DIMACS challenge [9]. Again, experiments were performed with values k = 1 and k = 2. Each experiment was tested for 20 independent runs. Results of the experiments can be found in Table 3-3, showing average time to solution, maximum best solution, minimum best solution, and average best solution. These results have been compared with the tabu search results, along with the benchmark instances of [3] in Table 3-4 and Table 3-5. Comparison values indicate average time to solution and maximum best values. Values in square brackets are the bounds for non-optimal terminations in the branch-and-cut instances from [3]. Values in parenthesis are new bounds. Experiments that did not find optimal values (or current lower bounds) are marked with an asterisk (*). The exchange local search algorithm performed well on most problems. Although most search times were comparable with tabu search, some instances proved to require 18

19 more processing time. The resulting values from the experiments are comparable with the tabu search method. Most instances found the optimal solution or known lower bound. For some instances, the exchange algorithm found better solutions that the tabu search algorithm, including a new bound for hamming10-4.clq with k = 2. Comparing the overall results with tabu search, the exchange algorithm proves to be more stable. Maximum best solutions for both methods are similar, but average best solutions differ significantly. The exchange algorithm was consistently better at finding solutions for many instances, as can be seen in the difference of minimum best values. Overall, the exchange local search method is a well-performing heuristic for the maximum k-plex problem. 19

20 1 Greedy Algorithm 2 S 0 is the inital k-plex 3 S BL is the subset of S where N(v i ) S = S k for v i S 4 S S 0 5 while S is not maximal do 6 for i = 1 to n do 7 if N(v i ) S S k + 1 then 8 if N(v i ) S BL = S BL then 9 pot(v i ) N(v i ) S 10 if pot(v i ) > pot best then 11 pot best pot(v i ) 12 v best v 13 end if 14 end if 15 end if 16 end for 17 S S v best 18 end while Figure 3-1. Greedy algorithm for k-plex 1 Tabu Algorithm 2 S 0 is the inital k-plex 3 S S 0 4 while time < run time do 5 while S is not maximal do 6 Select v i S using the greedy algorithm; 7 if v i is not tabu then 8 S S v i 9 end if 10 end while 11 if S > S best then 12 S best S 13 end if 14 Select a random v i from S; 15 S S \ v i 16 end while Figure 3-2. Tabu algorithm for k-plex 20

21 1 Exchange Algorithm 2 S 0 is the inital k-plex 3 S S 0 4 while time < run time do 5 while S is not maximal do 6 Select v i S using the greedy algorithm; 7 S S v i 8 end while 9 Select a random v new from S; 10 while an exchange partner is not selected do 11 Select a random exchange partner v old from S; 12 if N(v new ) S BL N(v old ) S BL then 13 ifv old N(v new ) then 14 if N(v new ) S S k + 1 then Select v old as exchange partner; S vold 17 S v new 18 end if 19 end if 20 else 21 if N(v new ) S S k then Select v old as exchange partner; S vold 24 S v new 25 end if 26 end else 27 end if 28 end while 29 end while Figure 3-3. Exchange algorithm for k-plex 21

22 Table 3-1. The tested DIMACS instances Graph V E c-fat200-1.clq c-fat200-2.clq c-fat200-5.clq c-fat500-1.clq c-fat500-2.clq c-fat500-5.clq c-fat clq hamming6-2.clq hamming6-4.clq hamming8-2.clq hamming8-4.clq hamming10-2.clq hamming10-4.clq johnson8-2-4.clq johnson8-4-4.clq MANN a9.clq MANN a27.clq MANN a45.clq san clq keller4.clq

23 Table 3-2. Tabu results for k = 1 and k = 2 k = 1 k = 2 Graph Time(s) Max Min Avg Time(s) Max Min Avg c-fat200-1.clq c-fat200-2.clq c-fat200-5.clq c-fat500-1.clq c-fat500-2.clq c-fat500-5.clq c-fat clq hamming6-2.clq hamming6-4.clq hamming8-2.clq hamming8-4.clq hamming10-2.clq hamming10-4.clq johnson8-2-4.clq johnson8-4-4.clq MANN a9.clq MANN a27.clq MANN a45.clq san clq keller4.clq

24 Table 3-3. Exchange results for k = 1 and k = 2 k = 1 k = 2 Graph Time(s) Max Min Avg Time(s) Max Min Avg c-fat200-1.clq c-fat200-2.clq c-fat200-5.clq c-fat500-1.clq c-fat500-2.clq c-fat500-5.clq c-fat clq hamming6-2.clq hamming6-4.clq hamming8-2.clq hamming8-4.clq hamming10-2.clq hamming10-4.clq johnson8-2-4.clq johnson8-4-4.clq MANN a9.clq MANN a27.clq MANN a45.clq san clq keller4.clq

25 Table 3-4. Result comparison for k = 1 Graph Tabu Time(s) Exg Time(s) BC Time(s) c-fat200-1.clq c-fat200-2.clq c-fat200-5.clq c-fat500-1.clq c-fat500-2.clq c-fat500-5.clq c-fat clq hamming6-2.clq hamming6-4.clq hamming8-2.clq hamming8-4.clq hamming10-2.clq hamming10-4.clq (40) (40) [38,379] johnson8-2-4.clq johnson8-4-4.clq MANN a9.clq MANN a27.clq (126) (126) [125,148] MANN a45.clq 334* * [342,422] san clq 12* 0.00 (18) [17,27] keller4.clq 9*

26 Table 3-5. Result comparison for k = 2 Graph Tabu Time(s) Exg Time(s) BC Time(s) c-fat200-1.clq c-fat200-2.clq c-fat200-5.clq c-fat500-1.clq c-fat500-2.clq c-fat500-5.clq c-fat clq hamming6-2.clq hamming6-4.clq hamming8-2.clq [128,130] hamming8-4.clq [16,80] hamming10-2.clq [512,534] hamming10-4.clq (47) (48) [45,458] johnson8-2-4.clq johnson8-4-4.clq [14,16] MANN a9.clq MANN a27.clq 235* * [236,260] MANN a45.clq 660* * [662,739] san clq 24* * 0.00 [62,86] keller4.clq 15* * 0.00 [40,45]

27 CHAPTER 4 CONCLUSIONS This study developed new methods for applying heuristics to the maximum k-plex problem. The first proposed algorithm used tabu search, along with a greedy technique, to implement a local search procedure. The tabu algorithm was tested on a set of problems provided by DIMACS [9]. It showed good results for most problems, finding optimal solutions for 15 of 16 known solutions for k = 1 and 11 of 11 known solutions for k = 2. For benchmarks with unknown optimal solutions, the algorithm found new bounds for two of four solutions with k = 1 and one of nine solutions with k = 2. A small number of instances did not break lower bounds. The exchange algorithm was experimented on the same set of DIMACS instances as tabu search [9]. In comparison with tabu search, the exchange algorithm performed more reliably; terminating solutions were consistently higher. For values of both k = 1 and k = 2, all known optimal solutions were found. For instances with bounded solutions, the exchange algorithm met or surpassed lower bounds for three of four solutions with k = 1 and five of nine solutions with k = 2. Again, a small number of instances did not meet lower bounds, and these examples of poor performance are primarily shown in high density graphs. It is often necessary to find solutions to optimization problems in a limited period of time. Development of approximation methods to estimate solutions quickly is necessary for the quick evaluation of large and complex problems. These heuristics provide new method that can be applied to finding k-plexes in a graph. Extending the research of local search methods for the maximum k-plex problem is very promising. With additional tuning, the tabu search algorithm could improve dramatically from its current performance. Incorporating techniques that can be adjusted during the search procedure for both the tabu and exchange algorithms is expected to 27

28 have improved results. In addition to building upon current search techniques, other methods, such as simulated annealing, may have successful results as with many other applications. Further research regarding the maximum k-plex problem will have positive effects on the field of optimization and social sciences. Applications of finding large k-plex values in graphs are expected to have an impact on determining cohesive subgroups in many real life situations. Other than social network analysis, the maximum k-plex problem has the opportunity and possibility of being implemented in many other fields of study. 28

29 REFERENCES [1] Abello J, Pardalos PM, Resende MGC. On maximum clique problems in very large graphs, External memory algorithms and visualization. DIMACS Series on Discrete Mathematics and Theoretical Computer Science. American Mathematical Society 1999;50: [2] Alba RD. A graph-theoretic definition of a sociometric clique. Journal of Mathematical Sociology 1973;3: [3] Balasundaram B, Butenko S, Hicks IV, Sachdeva S. Clique relaxations in social network analysis: The maximum k-plex problem. Submitted; [4] Battiti R, Tecchiolli G. The Reactive Tabu Search. ORSA Journal on Computing 1994;6: [5] Bellare M, Goldreich O, Sudan M. Free Bits, PCPs, and NonapproximabilityTowards Tight Results. SIAM Journal on Computing 1998;27(3): [6] Bomze IM, Budinich M, Pardalos PM, Pelillo M. The maximum clique problem. Handbook of Combinatorial Optimization, Kluwer Academic Publishers, Dordrecht, The Netherlands 1999;1:1-74. [7] DellAmico M, Trubian M. Applying tabu search to the job-shop scheduling problem. Annals of Operations Research 1993;41: [8] Deo N. Graph theory with applications to engineering and computer science. Englewood Cliffs, NJ: Prentice-Hall; [9] DIMACS. Cliques, Coloring, and Satisability: Second DIMACS Implementation Challenge Accessed Februrary [10] Freeman LC. The sociological concept of a group : An empirical test of two models. American Journal of Sociology 1992; 98: [11] Gendreau M, Laporte G, Musaraganyi C, Taillard ED. A tabu search heuristic for the heterogeneous fleet vehicle routing problem. Computers and Operations Research 1999;26: [12] Glover F, Laguna M. Tabu Search. Handbook of Combinatorial Optimization, Kluwer Academic Publishers, Dordrecht, The Netherlands 1999;3: [13] Glover F. Tabu Search Fundamentals and Uses. University of Colorado, Boulder. June Accessed March [14] Glover F. Tabu search: Part I. ORSA Journal on Computing 1989;1(3): [15] Glover F. Tabu search: Part II. ORSA Journal on Computing 1990;2(1):

30 [16] Hertz A, de Werra D. Using tabu search techniques for graph coloring. Computing 1987;39: [17] Hubscher R, Glover F. Applying tabu search with influential diversification to multiprocessor scheduling. Computers and Operations Research 1994;21(8): [18] Luce RD. Connectivity and generalized cliques in sociometric group structure. Psychometrika 1950;15: [19] Mokken RJ. Cliques, clubs and clans. Quality and Quantity 1979;13: [20] Osman IH, Laporte G. Metaheuristics: a bibliography. Annals of Operations Research 1996;63: [21] Pardalos PM, Xue J. The maximum clique problem. Journal of Global Optimization 1994;4(3): [22] Scott J. Social Network Analysis: A Handbook. Sage Publications, London, 2nd edition; [23] Seidman SB, Foster BL. A graph theoretic generalization of the clique concept. Journal of Mathematical Sociology 1978;6: [24] Tang H, Miller-Hooks E. A tabu search heuristic for the team orienteering problem. Computers and Operations Research 2005;32: [25] Wasserman S, Faust K. Social Network Analysis. Cambridge University Press;

31 BIOGRAPHICAL SKETCH The author, Erika Short, is a master s student at the University of Florida. She studies in the department of Industrial and Systems Engineering, with her primary focus in operations research. 31

Variable Objective Search

Variable Objective Search Variable Objective Search Sergiy Butenko, Oleksandra Yezerska, and Balabhaskar Balasundaram Abstract This paper introduces the variable objective search framework for combinatorial optimization. The method

More information

Combinatorial Algorithms for the Maximum k-plex Problem

Combinatorial Algorithms for the Maximum k-plex Problem Combinatorial Algorithms for the Maximum k-plex Problem Benjamin McClosky, Illya V. Hicks Department of Computational and Applied Mathematics, Rice University, 6100 Main St - MS 134, Houston, Texas 77005-1892,

More information

On a Polynomial Fractional Formulation for Independence Number of a Graph

On a Polynomial Fractional Formulation for Independence Number of a Graph On a Polynomial Fractional Formulation for Independence Number of a Graph Balabhaskar Balasundaram and Sergiy Butenko Department of Industrial Engineering, Texas A&M University, College Station, Texas

More information

Combinatorial algorithms for the maximum k-plex problem

Combinatorial algorithms for the maximum k-plex problem DOI 10.1007/s10878-010-9338-2 Combinatorial algorithms for the maximum k-plex problem Benjamin McClosky Illya V. Hicks Springer Science+Business Media, LLC 2010 Abstract The maximum clique problem provides

More information

Peripheries of Cohesive Subsets

Peripheries of Cohesive Subsets Peripheries of Cohesive Subsets Martin G. Everett University of Greenwich School of Computing and Mathematical Sciences 30 Park Row London SE10 9LS Tel: (0181) 331-8716 Fax: (181) 331-8665 E-mail: m.g.everett@gre.ac.uk

More information

FINDING INDEPENDENT SETS IN A GRAPH USING CONTINUOUS MULTIVARIABLE POLYNOMIAL FORMULATIONS

FINDING INDEPENDENT SETS IN A GRAPH USING CONTINUOUS MULTIVARIABLE POLYNOMIAL FORMULATIONS FINDING INDEPENDENT SETS IN A GRAPH USING CONTINUOUS MULTIVARIABLE POLYNOMIAL FORMULATIONS J. ABELLO, S. BUTENKO, P.M. PARDALOS, AND M.G.C. RESENDE Abstract. Two continuous formulations of the maximum

More information

A SEQUENTIAL ELIMINATION ALGORITHM FOR COMPUTING BOUNDS ON THE CLIQUE NUMBER OF A GRAPH

A SEQUENTIAL ELIMINATION ALGORITHM FOR COMPUTING BOUNDS ON THE CLIQUE NUMBER OF A GRAPH A SEQUENTIAL ELIMINATION ALGORITHM FOR COMPUTING BOUNDS ON THE CLIQUE NUMBER OF A GRAPH Bernard Gendron Département d informatique et de recherche opérationnelle and Centre de recherche sur les transports

More information

Peripheries of cohesive subsets

Peripheries of cohesive subsets Ž. Social Networks 21 1999 397 407 www.elsevier.comrlocatersocnet Peripheries of cohesive subsets Martin G. Everett a,), Stephen P. Borgatti b,1 a UniÕersity of Greenwich School of Computing and Mathematical

More information

Tabu Search. Biological inspiration is memory the ability to use past experiences to improve current decision making.

Tabu Search. Biological inspiration is memory the ability to use past experiences to improve current decision making. Tabu Search Developed by Fred Glover in the 1970 s. Dr Glover is a business professor at University of Colorado at Boulder. Developed specifically as a combinatorial optimization tool. Biological inspiration

More information

Single Solution-based Metaheuristics

Single Solution-based Metaheuristics Parallel Cooperative Optimization Research Group Single Solution-based Metaheuristics E-G. Talbi Laboratoire d Informatique Fondamentale de Lille Single solution-based metaheuristics Improvement of a solution.

More information

MAX-2-SAT: How Good is Tabu Search in the Worst-Case?

MAX-2-SAT: How Good is Tabu Search in the Worst-Case? MAX-2-SAT: How Good is Tabu Search in the Worst-Case? Monaldo Mastrolilli IDSIA Galleria 2, 6928 Manno, Switzerland monaldo@idsia.ch Luca Maria Gambardella IDSIA Galleria 2, 6928 Manno, Switzerland luca@idsia.ch

More information

Metaheuristics and Local Search

Metaheuristics and Local Search Metaheuristics and Local Search 8000 Discrete optimization problems Variables x 1,..., x n. Variable domains D 1,..., D n, with D j Z. Constraints C 1,..., C m, with C i D 1 D n. Objective function f :

More information

Novel Approaches for Analyzing Biological Networks

Novel Approaches for Analyzing Biological Networks Novel Approaches for Analyzing Biological Networks Balabhaskar Balasundaram, Sergiy Butenko and Svyatoslav Trukhanov Department of Industrial Engineering, Texas A&M University, College Station, Texas 77843,

More information

A pruning pattern list approach to the permutation flowshop scheduling problem

A pruning pattern list approach to the permutation flowshop scheduling problem A pruning pattern list approach to the permutation flowshop scheduling problem Takeshi Yamada NTT Communication Science Laboratories, 2-4 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, JAPAN E-mail :

More information

Algorithms and Complexity theory

Algorithms and Complexity theory Algorithms and Complexity theory Thibaut Barthelemy Some slides kindly provided by Fabien Tricoire University of Vienna WS 2014 Outline 1 Algorithms Overview How to write an algorithm 2 Complexity theory

More information

An Exact Algorithm for the Steiner Tree Problem with Delays

An Exact Algorithm for the Steiner Tree Problem with Delays Electronic Notes in Discrete Mathematics 36 (2010) 223 230 www.elsevier.com/locate/endm An Exact Algorithm for the Steiner Tree Problem with Delays Valeria Leggieri 1 Dipartimento di Matematica, Università

More information

Metaheuristics and Local Search. Discrete optimization problems. Solution approaches

Metaheuristics and Local Search. Discrete optimization problems. Solution approaches Discrete Mathematics for Bioinformatics WS 07/08, G. W. Klau, 31. Januar 2008, 11:55 1 Metaheuristics and Local Search Discrete optimization problems Variables x 1,...,x n. Variable domains D 1,...,D n,

More information

ROBUST CONNECTIVITY ISSUES IN DYNAMIC SENSOR NETWORKS FOR AREA SURVEILLANCE UNDER UNCERTAINTY

ROBUST CONNECTIVITY ISSUES IN DYNAMIC SENSOR NETWORKS FOR AREA SURVEILLANCE UNDER UNCERTAINTY ROBUST CONNECTIVITY ISSUES IN DYNAMIC SENSOR NETWORKS FOR AREA SURVEILLANCE UNDER UNCERTAINTY Konstantin Kalinchenko, Alexander Veremyev, Vladimir Boginski, David E. Jeffcoat and Stan Uryasev Abstract:

More information

An approach for the Class/Teacher Timetabling Problem using Graph Coloring

An approach for the Class/Teacher Timetabling Problem using Graph Coloring An approach for the Class/Teacher Timetabling Problem using Graph Coloring G. S. Bello M. C. Rangel M. C. S. Boeres Received: date / Accepted: date Keywords Timetabling Graph Coloring Metaheuristics Tabu

More information

A new ILS algorithm for parallel machine scheduling problems

A new ILS algorithm for parallel machine scheduling problems J Intell Manuf (2006) 17:609 619 DOI 10.1007/s10845-006-0032-2 A new ILS algorithm for parallel machine scheduling problems Lixin Tang Jiaxiang Luo Received: April 2005 / Accepted: January 2006 Springer

More information

Integer Programming Formulations for the Minimum Weighted Maximal Matching Problem

Integer Programming Formulations for the Minimum Weighted Maximal Matching Problem Optimization Letters manuscript No. (will be inserted by the editor) Integer Programming Formulations for the Minimum Weighted Maximal Matching Problem Z. Caner Taşkın Tınaz Ekim Received: date / Accepted:

More information

A tabu search algorithm for the open shop scheduling problem

A tabu search algorithm for the open shop scheduling problem A tabu search algorithm for the open shop scheduling problem 註解 [LEO1]: Ching-fang Liaw Department of Industrial Engineering and Management Chaoyang University of Technology Taiwan, R.O.C. August 11, 1997

More information

3.4 Relaxations and bounds

3.4 Relaxations and bounds 3.4 Relaxations and bounds Consider a generic Discrete Optimization problem z = min{c(x) : x X} with an optimal solution x X. In general, the algorithms generate not only a decreasing sequence of upper

More information

The combinatorics of pivoting for the maximum weight clique

The combinatorics of pivoting for the maximum weight clique Operations Research Letters 32 (2004) 523 529 Operations Research Letters wwwelseviercom/locate/dsw The combinatorics of pivoting for the maximum weight clique Marco Locatelli a; ;1, Immanuel M Bomze b,

More information

Hill climbing: Simulated annealing and Tabu search

Hill climbing: Simulated annealing and Tabu search Hill climbing: Simulated annealing and Tabu search Heuristic algorithms Giovanni Righini University of Milan Department of Computer Science (Crema) Hill climbing Instead of repeating local search, it is

More information

GRASP with path-relinking for the weighted maximum satisfiability problem

GRASP with path-relinking for the weighted maximum satisfiability problem GRASP with path-relinking for the weighted maximum satisfiability problem Paola Festa 1, Panos M. Pardalos 2, Leonidas S. Pitsoulis 3, and Mauricio G. C. Resende 4 1 Department of Mathematics and Applications,

More information

A new set of hybrid heuristics for the quadratic assignment problem

A new set of hybrid heuristics for the quadratic assignment problem A new set of hybrid heuristics for the quadratic assignment problem André de São Thiago Moreira Valdir Agustinho de Melo Paulo Oswaldo Boaventura Netto Programa de Engenharia de Produção, COPPE/UFRJ CP

More information

Design and Analysis of Algorithms

Design and Analysis of Algorithms CSE 0, Winter 08 Design and Analysis of Algorithms Lecture 8: Consolidation # (DP, Greed, NP-C, Flow) Class URL: http://vlsicad.ucsd.edu/courses/cse0-w8/ Followup on IGO, Annealing Iterative Global Optimization

More information

Lecture H2. Heuristic Methods: Iterated Local Search, Simulated Annealing and Tabu Search. Saeed Bastani

Lecture H2. Heuristic Methods: Iterated Local Search, Simulated Annealing and Tabu Search. Saeed Bastani Simulation Lecture H2 Heuristic Methods: Iterated Local Search, Simulated Annealing and Tabu Search Saeed Bastani saeed.bastani@eit.lth.se Spring 2017 Thanks to Prof. Arne Løkketangen at Molde University

More information

An Integrated Column Generation and Lagrangian Relaxation for Flowshop Scheduling Problems

An Integrated Column Generation and Lagrangian Relaxation for Flowshop Scheduling Problems Proceedings of the 2009 IEEE International Conference on Systems, Man, and Cybernetics San Antonio, TX, USA - October 2009 An Integrated Column Generation and Lagrangian Relaxation for Flowshop Scheduling

More information

Flow Shop and Job Shop Models

Flow Shop and Job Shop Models Outline DM87 SCHEDULING, TIMETABLING AND ROUTING Lecture 11 Flow Shop and Job Shop Models 1. Flow Shop 2. Job Shop Marco Chiarandini DM87 Scheduling, Timetabling and Routing 2 Outline Resume Permutation

More information

Meta-heuristics for combinatorial optimisation

Meta-heuristics for combinatorial optimisation Meta-heuristics for combinatorial optimisation João Pedro Pedroso Centro de Investigação Operacional Faculdade de Ciências da Universidade de Lisboa and Departamento de Ciência de Computadores Faculdade

More information

CS 583: Approximation Algorithms: Introduction

CS 583: Approximation Algorithms: Introduction CS 583: Approximation Algorithms: Introduction Chandra Chekuri January 15, 2018 1 Introduction Course Objectives 1. To appreciate that not all intractable problems are the same. NP optimization problems,

More information

Working Paper Series HEARIN CENTER FOR ENTERPRISE SCIENCE

Working Paper Series HEARIN CENTER FOR ENTERPRISE SCIENCE HCES -05-04 Working Paper Series HEARIN CENTER FOR ENTERPRISE SCIENCE Solving the Maximum Edge Weight Clique Problem via Unconstrained Quadratic Programming By Gary Kochenberger Fred Glover Bahram Alidaee

More information

An Optimization-Based Heuristic for the Split Delivery Vehicle Routing Problem

An Optimization-Based Heuristic for the Split Delivery Vehicle Routing Problem An Optimization-Based Heuristic for the Split Delivery Vehicle Routing Problem Claudia Archetti (1) Martin W.P. Savelsbergh (2) M. Grazia Speranza (1) (1) University of Brescia, Department of Quantitative

More information

New Lower Bounds on the Stability Number of a Graph

New Lower Bounds on the Stability Number of a Graph New Lower Bounds on the Stability Number of a Graph E. Alper Yıldırım June 27, 2007 Abstract Given a simple, undirected graph G, Motzkin and Straus [Canadian Journal of Mathematics, 17 (1965), 533 540]

More information

PROBLEM SOLVING AND SEARCH IN ARTIFICIAL INTELLIGENCE

PROBLEM SOLVING AND SEARCH IN ARTIFICIAL INTELLIGENCE Artificial Intelligence, Computational Logic PROBLEM SOLVING AND SEARCH IN ARTIFICIAL INTELLIGENCE Lecture 4 Metaheuristic Algorithms Sarah Gaggl Dresden, 5th May 2017 Agenda 1 Introduction 2 Constraint

More information

A new trust region technique for the maximum weight clique problem

A new trust region technique for the maximum weight clique problem A new trust region technique for the maximum weight clique problem Stanislav Busygin Industrial and Systems Engineering Department, University of Florida, 303 Weil Hall, Gainesville, FL 32611, USA Abstract

More information

Adaptive Dynamic Cost Updating Procedure for Solving Fixed Charge Network Flow Problems.

Adaptive Dynamic Cost Updating Procedure for Solving Fixed Charge Network Flow Problems. Adaptive Dynamic Cost Updating Procedure for Solving Fixed Charge Network Flow Problems. Artyom Nahapetyan, Panos Pardalos Center for Applied Optimization Industrial and Systems Engineering Department

More information

Finding Cohesive Subgroups in Social Networks

Finding Cohesive Subgroups in Social Networks Finding Cohesive Subgroups in Social Networks Illya V. Hicks Computational and Applied Mathematics Rice University RUSMP Fall Networking Conference September 19, 2009 Outline I. Basic Definitions II. III.

More information

Co-k-plex vertex partitions

Co-k-plex vertex partitions Co-k-plex vertex partitions Benjamin McClosky John D. Arellano Illya V. Hicks November 7, 2012 Abstract This paper studies co-k-plex vertex partitions and more specifically co-2-plex vertex partitions.

More information

Solving the maximum edge weight clique problem via unconstrained quadratic programming

Solving the maximum edge weight clique problem via unconstrained quadratic programming European Journal of Operational Research 181 (2007) 592 597 Discrete Optimization Solving the maximum edge weight clique problem via unconstrained quadratic programming Bahram Alidaee a, Fred Glover b,

More information

Safe Lower Bounds For Graph Coloring

Safe Lower Bounds For Graph Coloring Safe Lower Bounds For Graph Coloring Stephan Held, Edward C. Sewell, William Cook November 8, 2010 Abstract The best known method for determining lower bounds on the vertex coloring number of a graph is

More information

A Mixed-Integer Linear Program for the Traveling Salesman Problem with Structured Time Windows

A Mixed-Integer Linear Program for the Traveling Salesman Problem with Structured Time Windows A Mixed-Integer Linear Program for the Traveling Salesman Problem with Structured Time Windows Philipp Hungerländer Christian Truden 5th January 2017 Abstract In this extended abstract we introduce the

More information

Cliques and Clubs.

Cliques and Clubs. Cliques and Clubs Petr A. Golovach 1, Pinar Heggernes 1, Dieter Kratsch 2, and Arash Rafiey 1 1 Department of Informatics, University of Bergen, Norway, {petr.golovach,pinar.heggernes,arash.rafiey}@ii.uib.no

More information

ICS 252 Introduction to Computer Design

ICS 252 Introduction to Computer Design ICS 252 fall 2006 Eli Bozorgzadeh Computer Science Department-UCI References and Copyright Textbooks referred [Mic94] G. De Micheli Synthesis and Optimization of Digital Circuits McGraw-Hill, 1994. [CLR90]

More information

f-flip strategies for unconstrained binary quadratic programming

f-flip strategies for unconstrained binary quadratic programming Ann Oper Res (2016 238:651 657 DOI 10.1007/s10479-015-2076-1 NOTE f-flip strategies for unconstrained binary quadratic programming Fred Glover 1 Jin-Kao Hao 2,3 Published online: 11 December 2015 Springer

More information

Vehicle Routing and Scheduling. Martin Savelsbergh The Logistics Institute Georgia Institute of Technology

Vehicle Routing and Scheduling. Martin Savelsbergh The Logistics Institute Georgia Institute of Technology Vehicle Routing and Scheduling Martin Savelsbergh The Logistics Institute Georgia Institute of Technology Vehicle Routing and Scheduling Part II: Algorithmic Enhancements Handling Practical Complexities

More information

COMPUTATIONAL COMPLEXITY

COMPUTATIONAL COMPLEXITY ATHEATICS: CONCEPTS, AND FOUNDATIONS Vol. III - Computational Complexity - Osamu Watanabe COPUTATIONAL COPLEXITY Osamu Watanabe Tokyo Institute of Technology, Tokyo, Japan Keywords: {deterministic, randomized,

More information

12. LOCAL SEARCH. gradient descent Metropolis algorithm Hopfield neural networks maximum cut Nash equilibria

12. LOCAL SEARCH. gradient descent Metropolis algorithm Hopfield neural networks maximum cut Nash equilibria 12. LOCAL SEARCH gradient descent Metropolis algorithm Hopfield neural networks maximum cut Nash equilibria Lecture slides by Kevin Wayne Copyright 2005 Pearson-Addison Wesley h ttp://www.cs.princeton.edu/~wayne/kleinberg-tardos

More information

A tabu search based memetic algorithm for the maximum diversity problem

A tabu search based memetic algorithm for the maximum diversity problem A tabu search based memetic algorithm for the maximum diversity problem Yang Wang a,b, Jin-Kao Hao b,, Fred Glover c, Zhipeng Lü d a School of Management, Northwestern Polytechnical University, 127 Youyi

More information

Zebo Peng Embedded Systems Laboratory IDA, Linköping University

Zebo Peng Embedded Systems Laboratory IDA, Linköping University TDTS 01 Lecture 8 Optimization Heuristics for Synthesis Zebo Peng Embedded Systems Laboratory IDA, Linköping University Lecture 8 Optimization problems Heuristic techniques Simulated annealing Genetic

More information

NP-complete problems. CSE 101: Design and Analysis of Algorithms Lecture 20

NP-complete problems. CSE 101: Design and Analysis of Algorithms Lecture 20 NP-complete problems CSE 101: Design and Analysis of Algorithms Lecture 20 CSE 101: Design and analysis of algorithms NP-complete problems Reading: Chapter 8 Homework 7 is due today, 11:59 PM Tomorrow

More information

Permutation distance measures for memetic algorithms with population management

Permutation distance measures for memetic algorithms with population management MIC2005: The Sixth Metaheuristics International Conference??-1 Permutation distance measures for memetic algorithms with population management Marc Sevaux Kenneth Sörensen University of Valenciennes, CNRS,

More information

A comparison of sequencing formulations in a constraint generation procedure for avionics scheduling

A comparison of sequencing formulations in a constraint generation procedure for avionics scheduling A comparison of sequencing formulations in a constraint generation procedure for avionics scheduling Department of Mathematics, Linköping University Jessika Boberg LiTH-MAT-EX 2017/18 SE Credits: Level:

More information

GRASP FOR LINEAR INTEGER PROGRAMMING

GRASP FOR LINEAR INTEGER PROGRAMMING Chapter 1 GRASP FOR LINEAR INTEGER PROGRAMMING Teresa Neto tneto@matestvipvpt João Pedro Pedroso jpp@nccuppt Departamento de Ciência de Computadores Faculdade de Ciências da Universidade do Porto Rua do

More information

The shortest path tour problem: problem definition, modeling, and optimization

The shortest path tour problem: problem definition, modeling, and optimization The shortest path tour problem: problem definition, modeling, and optimization Paola Festa Department of Mathematics and Applications, University of apoli FEDERICO II Compl. MSA, Via Cintia, 86 apoli,

More information

Axioms of Density: How to Define and Detect the Densest Subgraph

Axioms of Density: How to Define and Detect the Densest Subgraph RT097 Computer Science; Network pages Research Report March 31, 016 Axioms of Density: How to Define and Detect the Densest Subgraph Hiroki Yanagisawa and Satoshi Hara IBM Research - Tokyo IBM Japan, Ltd.

More information

Finding independent sets in a graph using continuous multivariable polynomial formulations

Finding independent sets in a graph using continuous multivariable polynomial formulations Journal of Global Optimization 21: 111 137, 2001. 2001 Kluwer Academic Publishers. Printed in the Netherlands. 111 Finding independent sets in a graph using continuous multivariable polynomial formulations

More information

Artificial Intelligence Heuristic Search Methods

Artificial Intelligence Heuristic Search Methods Artificial Intelligence Heuristic Search Methods Chung-Ang University, Jaesung Lee The original version of this content is created by School of Mathematics, University of Birmingham professor Sandor Zoltan

More information

Lin-Kernighan Heuristic. Simulated Annealing

Lin-Kernighan Heuristic. Simulated Annealing DM63 HEURISTICS FOR COMBINATORIAL OPTIMIZATION Lecture 6 Lin-Kernighan Heuristic. Simulated Annealing Marco Chiarandini Outline 1. Competition 2. Variable Depth Search 3. Simulated Annealing DM63 Heuristics

More information

Analysis of a Large Structure/Biological Activity. Data Set Using Recursive Partitioning and. Simulated Annealing

Analysis of a Large Structure/Biological Activity. Data Set Using Recursive Partitioning and. Simulated Annealing Analysis of a Large Structure/Biological Activity Data Set Using Recursive Partitioning and Simulated Annealing Student: Ke Zhang MBMA Committee: Dr. Charles E. Smith (Chair) Dr. Jacqueline M. Hughes-Oliver

More information

A.I.: Beyond Classical Search

A.I.: Beyond Classical Search A.I.: Beyond Classical Search Random Sampling Trivial Algorithms Generate a state randomly Random Walk Randomly pick a neighbor of the current state Both algorithms asymptotically complete. Overview Previously

More information

Solving Elementary Shortest-Path Problems as Mixed-Integer Programs

Solving Elementary Shortest-Path Problems as Mixed-Integer Programs Gutenberg School of Management and Economics Discussion Paper Series Solving Elementary Shortest-Path Problems as Mixed-Integer Programs Michael Drexl and Stefan Irnich Januar 2012 Discussion paper number

More information

THE CO-2-PLEX POLYTOPE AND INTEGRAL SYSTEMS

THE CO-2-PLEX POLYTOPE AND INTEGRAL SYSTEMS THE CO-2-PLEX POLYTOPE AND INTEGRAL SYSTEMS BENJAMIN MCCLOSKY AND ILLYA V. HICKS Abstract. k-plexes are cohesive subgraphs which were introduced to relax the structure of cliques. A co-k-plex is the complement

More information

Recoverable Robustness in Scheduling Problems

Recoverable Robustness in Scheduling Problems Master Thesis Computing Science Recoverable Robustness in Scheduling Problems Author: J.M.J. Stoef (3470997) J.M.J.Stoef@uu.nl Supervisors: dr. J.A. Hoogeveen J.A.Hoogeveen@uu.nl dr. ir. J.M. van den Akker

More information

Complexity of local search for the p-median problem

Complexity of local search for the p-median problem Complexity of local search for the p-median problem Ekaterina Alekseeva, Yuri Kochetov, Alexander Plyasunov Sobolev Institute of Mathematics, Novosibirsk {ekaterina2, jkochet, apljas}@math.nsc.ru Abstract

More information

A tabu search algorithm for the minmax regret minimum spanning tree problem with interval data

A tabu search algorithm for the minmax regret minimum spanning tree problem with interval data Noname manuscript No. (will be inserted by the editor) A tabu search algorithm for the minmax regret minimum spanning tree problem with interval data Adam Kasperski Mariusz Makuchowski Pawe l Zieliński

More information

Groups of vertices and Core-periphery structure. By: Ralucca Gera, Applied math department, Naval Postgraduate School Monterey, CA, USA

Groups of vertices and Core-periphery structure. By: Ralucca Gera, Applied math department, Naval Postgraduate School Monterey, CA, USA Groups of vertices and Core-periphery structure By: Ralucca Gera, Applied math department, Naval Postgraduate School Monterey, CA, USA Mostly observed real networks have: Why? Heavy tail (powerlaw most

More information

Minimum Linear Arrangements

Minimum Linear Arrangements Minimum Linear Arrangements Rafael Andrade, Tibérius Bonates, Manoel Câmpelo, Mardson Ferreira ParGO - Research team in Parallel computing, Graph theory and Optimization Department of Statistics and Applied

More information

The two-machine flowshop total completion time problem: A branch-and-bound based on network-flow formulation

The two-machine flowshop total completion time problem: A branch-and-bound based on network-flow formulation The two-machine flowshop total completion time problem: A branch-and-bound based on network-flow formulation Boris Detienne 1, Ruslan Sadykov 1, Shunji Tanaka 2 1 : Team Inria RealOpt, University of Bordeaux,

More information

Algorithm Design Strategies V

Algorithm Design Strategies V Algorithm Design Strategies V Joaquim Madeira Version 0.0 October 2016 U. Aveiro, October 2016 1 Overview The 0-1 Knapsack Problem Revisited The Fractional Knapsack Problem Greedy Algorithms Example Coin

More information

A Fast Heuristic for GO and MINLP

A Fast Heuristic for GO and MINLP A Fast Heuristic for GO and MINLP John W. Chinneck, M. Shafique, Systems and Computer Engineering Carleton University, Ottawa, Canada Introduction Goal: Find a good quality GO/MINLP solution quickly. Trade

More information

5. Simulated Annealing 5.1 Basic Concepts. Fall 2010 Instructor: Dr. Masoud Yaghini

5. Simulated Annealing 5.1 Basic Concepts. Fall 2010 Instructor: Dr. Masoud Yaghini 5. Simulated Annealing 5.1 Basic Concepts Fall 2010 Instructor: Dr. Masoud Yaghini Outline Introduction Real Annealing and Simulated Annealing Metropolis Algorithm Template of SA A Simple Example References

More information

Ant Colony Optimization: an introduction. Daniel Chivilikhin

Ant Colony Optimization: an introduction. Daniel Chivilikhin Ant Colony Optimization: an introduction Daniel Chivilikhin 03.04.2013 Outline 1. Biological inspiration of ACO 2. Solving NP-hard combinatorial problems 3. The ACO metaheuristic 4. ACO for the Traveling

More information

Local Search & Optimization

Local Search & Optimization Local Search & Optimization CE417: Introduction to Artificial Intelligence Sharif University of Technology Spring 2018 Soleymani Artificial Intelligence: A Modern Approach, 3 rd Edition, Chapter 4 Some

More information

Office of Naval Research contract no. N K-0377 (9) FINAL TECHNICAL REPORT

Office of Naval Research contract no. N K-0377 (9) FINAL TECHNICAL REPORT Office of Naval Research contract no. N00014-88-K-0377 (9) The Concept of Best Probability in the Analysis of Approximation Algorithms 0) - Dorit S. Hochbaum, Principal Investigator oom FINAL TECHNICAL

More information

Decision Diagrams for Discrete Optimization

Decision Diagrams for Discrete Optimization Decision Diagrams for Discrete Optimization Willem Jan van Hoeve Tepper School of Business Carnegie Mellon University www.andrew.cmu.edu/user/vanhoeve/mdd/ Acknowledgments: David Bergman, Andre Cire, Samid

More information

AN ANT APPROACH FOR STRUCTURED QUADRATIC ASSIGNMENT PROBLEMS

AN ANT APPROACH FOR STRUCTURED QUADRATIC ASSIGNMENT PROBLEMS AN ANT APPROACH FOR STRUCTURED QUADRATIC ASSIGNMENT PROBLEMS Éric D. Taillard, Luca M. Gambardella IDSIA, Corso Elvezia 36, CH-6900 Lugano, Switzerland. Extended abstract IDSIA-22-97 ABSTRACT. The paper

More information

Fixed interval scheduling problems under uncertainty

Fixed interval scheduling problems under uncertainty Fixed interval scheduling problems under uncertainty Martin Branda et al. Charles University Faculty of Mathematics and Physics Department of Probability and Mathematical Statistics Stochastic Programming

More information

Modular Monochromatic Colorings, Spectra and Frames in Graphs

Modular Monochromatic Colorings, Spectra and Frames in Graphs Western Michigan University ScholarWorks at WMU Dissertations Graduate College 12-2014 Modular Monochromatic Colorings, Spectra and Frames in Graphs Chira Lumduanhom Western Michigan University, chira@swu.ac.th

More information

An improved approximation algorithm for the stable marriage problem with one-sided ties

An improved approximation algorithm for the stable marriage problem with one-sided ties Noname manuscript No. (will be inserted by the editor) An improved approximation algorithm for the stable marriage problem with one-sided ties Chien-Chung Huang Telikepalli Kavitha Received: date / Accepted:

More information

Dominating Set. Chapter 26

Dominating Set. Chapter 26 Chapter 26 Dominating Set In this chapter we present another randomized algorithm that demonstrates the power of randomization to break symmetries. We study the problem of finding a small dominating set

More information

Dominating Set. Chapter Sequential Greedy Algorithm 294 CHAPTER 26. DOMINATING SET

Dominating Set. Chapter Sequential Greedy Algorithm 294 CHAPTER 26. DOMINATING SET 294 CHAPTER 26. DOMINATING SET 26.1 Sequential Greedy Algorithm Chapter 26 Dominating Set Intuitively, to end up with a small dominating set S, nodes in S need to cover as many neighbors as possible. It

More information

Lecture 9 Evolutionary Computation: Genetic algorithms

Lecture 9 Evolutionary Computation: Genetic algorithms Lecture 9 Evolutionary Computation: Genetic algorithms Introduction, or can evolution be intelligent? Simulation of natural evolution Genetic algorithms Case study: maintenance scheduling with genetic

More information

Enumeration and symmetry of edit metric spaces. Jessie Katherine Campbell. A dissertation submitted to the graduate faculty

Enumeration and symmetry of edit metric spaces. Jessie Katherine Campbell. A dissertation submitted to the graduate faculty Enumeration and symmetry of edit metric spaces by Jessie Katherine Campbell A dissertation submitted to the graduate faculty in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY

More information

Project in Computational Game Theory: Communities in Social Networks

Project in Computational Game Theory: Communities in Social Networks Project in Computational Game Theory: Communities in Social Networks Eldad Rubinstein November 11, 2012 1 Presentation of the Original Paper 1.1 Introduction In this section I present the article [1].

More information

Chapter 34: NP-Completeness

Chapter 34: NP-Completeness Graph Algorithms - Spring 2011 Set 17. Lecturer: Huilan Chang Reference: Cormen, Leiserson, Rivest, and Stein, Introduction to Algorithms, 2nd Edition, The MIT Press. Chapter 34: NP-Completeness 2. Polynomial-time

More information

Introduction to Social Network Analysis PSU Quantitative Methods Seminar, June 15

Introduction to Social Network Analysis PSU Quantitative Methods Seminar, June 15 Introduction to Social Network Analysis PSU Quantitative Methods Seminar, June 15 Jeffrey A. Smith University of Nebraska-Lincoln Department of Sociology Course Website https://sites.google.com/site/socjasmith/teaching2/psu_social_networks_seminar

More information

A parallel metaheuristics for the single machine total weighted tardiness problem with sequence-dependent setup times

A parallel metaheuristics for the single machine total weighted tardiness problem with sequence-dependent setup times A parallel metaheuristics for the single machine total weighted tardiness problem with sequence-dependent setup times Wojciech Bożejko Wroc law University of Technology Institute of Computer Engineering,

More information

HETEROGENEOUS QUANTUM COMPUTING FOR SATELLITE OPTIMIZATION

HETEROGENEOUS QUANTUM COMPUTING FOR SATELLITE OPTIMIZATION HETEROGENEOUS QUANTUM COMPUTING FOR SATELLITE OPTIMIZATION GIDEON BASS BOOZ ALLEN HAMILTON September 2017 COLLABORATORS AND PARTNERS Special thanks to: Brad Lackey (UMD/QuICS) for advice and suggestions

More information

Matteo Fischetti, DEI, University of Padova. COFIN Matheon workshop, Villa Vigoni (Como), May 2006

Matteo Fischetti, DEI, University of Padova. COFIN Matheon workshop, Villa Vigoni (Como), May 2006 Matteo Fischetti, DEI, University of Padova COFIN Matheon workshop, Villa Vigoni (Como), May 2006 1 MIP solvers for hard optimization problems Mixed-integer linear programming (MIP) plays a central role

More information

CS 781 Lecture 9 March 10, 2011 Topics: Local Search and Optimization Metropolis Algorithm Greedy Optimization Hopfield Networks Max Cut Problem Nash

CS 781 Lecture 9 March 10, 2011 Topics: Local Search and Optimization Metropolis Algorithm Greedy Optimization Hopfield Networks Max Cut Problem Nash CS 781 Lecture 9 March 10, 2011 Topics: Local Search and Optimization Metropolis Algorithm Greedy Optimization Hopfield Networks Max Cut Problem Nash Equilibrium Price of Stability Coping With NP-Hardness

More information

Disconnecting Networks via Node Deletions

Disconnecting Networks via Node Deletions 1 / 27 Disconnecting Networks via Node Deletions Exact Interdiction Models and Algorithms Siqian Shen 1 J. Cole Smith 2 R. Goli 2 1 IOE, University of Michigan 2 ISE, University of Florida 2012 INFORMS

More information

5. Simulated Annealing 5.2 Advanced Concepts. Fall 2010 Instructor: Dr. Masoud Yaghini

5. Simulated Annealing 5.2 Advanced Concepts. Fall 2010 Instructor: Dr. Masoud Yaghini 5. Simulated Annealing 5.2 Advanced Concepts Fall 2010 Instructor: Dr. Masoud Yaghini Outline Acceptance Function Initial Temperature Equilibrium State Cooling Schedule Stopping Condition Handling Constraints

More information

CHARACTERIZING STRUCTURALLY COHESIVE CLUSTERS IN NETWORKS: THEORY AND ALGORITHMS. A Dissertation CHITRA BALASUBRAMANIAM

CHARACTERIZING STRUCTURALLY COHESIVE CLUSTERS IN NETWORKS: THEORY AND ALGORITHMS. A Dissertation CHITRA BALASUBRAMANIAM CHARACTERIZING STRUCTURALLY COHESIVE CLUSTERS IN NETWORKS: THEORY AND ALGORITHMS A Dissertation by CHITRA BALASUBRAMANIAM Submitted to the Office of Graduate and Professional Studies of Texas A&M University

More information

IBM Research Report. Hilbert's Nullstellensatz and an Algorithm for Proving Combinatorial Infeasibility

IBM Research Report. Hilbert's Nullstellensatz and an Algorithm for Proving Combinatorial Infeasibility RC24472 (W0801-088) January 22, 2008 Mathematics IBM Research Report Hilbert's Nullstellensatz and an Algorithm for Proving Combinatorial Infeasibility J. A. De Loera 1, Jon Lee 2, P. Malkin 1, S. Margulies

More information

An Analysis on Recombination in Multi-Objective Evolutionary Optimization

An Analysis on Recombination in Multi-Objective Evolutionary Optimization An Analysis on Recombination in Multi-Objective Evolutionary Optimization Chao Qian, Yang Yu, Zhi-Hua Zhou National Key Laboratory for Novel Software Technology Nanjing University, Nanjing 20023, China

More information

More on NP and Reductions

More on NP and Reductions Indian Institute of Information Technology Design and Manufacturing, Kancheepuram Chennai 600 127, India An Autonomous Institute under MHRD, Govt of India http://www.iiitdm.ac.in COM 501 Advanced Data

More information

12. LOCAL SEARCH. gradient descent Metropolis algorithm Hopfield neural networks maximum cut Nash equilibria

12. LOCAL SEARCH. gradient descent Metropolis algorithm Hopfield neural networks maximum cut Nash equilibria Coping With NP-hardness Q. Suppose I need to solve an NP-hard problem. What should I do? A. Theory says you re unlikely to find poly-time algorithm. Must sacrifice one of three desired features. Solve

More information