A Hybrid Constraint Programming and Semidefinite Programming Approach for the Stable Set Problem

Many important NP-hard combinatorial problems can be efficiently approximated using semidefinite programming relaxations. We propose a new hierarchy of semidefinite relaxations for classes of such problems that are based on graphs and for which the projection of the problem onto a subgraph shares the same structure as the original problem. This includes the well-studied max-cut and stable-set problems. Each level k of the proposed hierarchy consists of the basic semidefinite relaxation of the problem augmented by the constraints enforcing the structural projection condition on every k-node subgraph of the problem. This hierarchy has the distinguishing feature that all the relaxations are formulated in the space of the original semidefinite relaxation. Because the size of the relaxations increases rapidly with the number of subgraphs, we explore the possibility of adding the projection constraints only for selected subgraphs. Preliminary computational results show that the proposed hierarchy yields improved bounds when compared to the initial relaxation for benchmark instances of the max-cut and stable-set problems, and that the improved bounds result in significantly smaller enumeration trees when the relaxation is used in a branch-and-bound scheme.

The exponential-time hypothesis (ETH) states that 3-SAT is not solvable in subexponential time, i.e. not solvable in O(c^n) time for arbitrary c > 1, where n denotes the number of variables. Problems like k-SAT can be viewed as special cases of the constraint satisfaction problem (CSP), which is the problem of determining whether a set of constraints is satisfiable. In this paper we study the worst-case time complexity of NP-complete CSPs. Our main interest is in the CSP problem parameterized by a constraint language Gamma (CSP(Gamma)), and how the choice of Gamma affects the time complexity. It is believed that CSP(Gamma) is either tractable or NP-complete, and the algebraic CSP dichotomy conjecture gives a sharp delineation of these two classes based on algebraic properties of constraint languages. Under this conjecture and the ETH, we first rule out the existence of subexponential algorithms for finite domain NP-complete CSP(Gamma) problems. This result also extends to certain infinite-domain CSPs and structurally restricted CSP(Gamma) problems. We then begin a study of the complexity of NP-complete CSPs where one is allowed to arbitrarily restrict the values of individual variables, which is a very well-studied subclass of CSPs. For such CSPs with finite domain D, we identify a relation SD such that (1) CSP({SD}) is NP-complete and (2) if CSP(Gamma) over D is NP-complete and solvable in O(c^n) time, then CSP({SD}) is solvable in O(c^n) time, too. Hence, the time complexity of CSP({SD}) is a lower bound for all CSPs of this particular kind. We also prove that the complexity of CSP({SD}) is decreasing when |D| increases, unless the ETH is false. This implies, for instance, that for every c>1 there exists a finite-domain Gamma such that CSP(Gamma) is NP complete and solvable in O(c^n) time.

Cell formation deals with grouping of machines and parts in manufacturing systems according to their compatibility. Manufacturing processes are surrounded with an abundance of complex constraints which should be considered carefully and represented clearly for obtaining high efficiency and productivity. Constraint programming is a new approach to combinatorial optimization and provides a rich language to represent complex constraints easily. However, the cell formation problems are well suited to be solved by constraint programming approach since the problem has many constraints such as part-machine requirements, availabilities in the system in terms of capacity, machine and worker abilities. In this study, the cell formation problem is modeled using machine, part processing and worker flexibilities via resource element–based representation. Resource elements define the processing requirements of parts and processing capabilities of machines and workers, which are resource-independent capability units. A total of 12 case problems are generated, and different search phases of constraint programming are defined for the solution procedure. The cell formation problem is modeled in both constraint programming and integer programming, and a comparative analysis of constraint programming and integer programming model solutions is done. The results indicate that both the models are effective and efficient in the solution of the cell formation problem.

In this paper a constraint programming (CP) approach for calculating release dates for lots within a supply chain environment is investigated. The lot start times are verified by a simulation model using different dispatching rules focusing on tardiness. To test the presented CP approach a simple fab model is constructed. The fab model consists of parallel batch machines as well as work centers and single machines. The investigated objectives are tardiness, earliness and cycle time. Due to the high complexity decomposition methods for the CP approach are tested. The results from the CP method are lot start dates which are verified by a downstream simulation run. The results show that the presented CP approach could outperform the simulation for all objectives. The content of this paper could be used as a first investigation for new scheduling methods within a supply chain management.

Constraint programming for solving various assembly line balancing problems

Large-neighbourhood search (LNS) improves an initial solution, hence it is not directly applicable to satisfaction problems. In order to use LNS in a constraint programming (CP) framework to solve satisfaction problems, we usually soften some hard-to-satisfy constraints by replacing them with penalty-function constraints. LNS is then used to reduce their penalty to zero, thus satisfying the original problem. However, this can give poor performance as the penalties rarely cause propagation and therefore do not drive each CP search, and by extension the LNS search, towards satisfying the replaced constraints until very late. Our key observation is that entirely replacing a constraint is often overkill, as the propagator for the replaced constraint could have performed some propagation without causing backtracking. We propose the notion of a non-failing propagator, which is subsumed just before causing a backtrack. We show that, by only making a few changes to an existing CP solver, any propagator can be made non-failing without modifying its code. Experimental evaluation shows that non-failing propagators, when used in conjunction with penalties, can greatly improve LNS performance compared to just having penalties. This allows us to leverage the power of the many sophisticated propagators that already exist in CP solvers, in order to use LNS for solving hard satisfaction problems and for finding initial solutions to hard-to-satisfy optimisation problems.

In this paper, we propose a new value ordering heuristic for weighted constraint satisfaction problems (WCSP) based on the quality of solutions. The H-quality of an assignment estimates its capacity at occurring in solutions of good quality. Experiments using limited discrepancy search on random WCSP instances and CELAR benchmarks show that our value ordering always outperforms MinAC in a significant way.

This paper investigates the use of abstract domains from Abstract Interpretation (AI) in the field of Constraint Programming (CP). CP solvers are generally very efficient on a specific constraint language, but can hardly be extended to tackle more general languages, both in terms of variable representation (discrete or continuous) and constraint type (global, arithmetic, etc.). For instance, linear constraints are usually solved with linear programming techniques, but non-linear ones have to be either linearized, reformulated or sent to an external solver. We approach this problem by adapting to CP a popular domain construction used to combine the power of several analyses in AI: the reduced product. We apply this product on two well-known abstract domains, Boxes and Polyhedra, that we lift to constraint solving. Finally we define general metrics for the quality of the solver results, and present a benchmark accordingly. Experiments show promising results and good performances.

The peptide-protein docking problem is an important problem in structural biology that facilitates rational and efficient drug design. In this work, we explore modeling and solving this problem with the quantum-amenable quadratic unconstrained binary optimization (QUBO) formalism. Our work extends recent efforts by incorporating the objectives and constraints associated with peptide cyclization and peptide-protein docking in the two-particle model on a tetrahedral lattice. We propose a “resource efficient” QUBO encoding for this problem, and baseline its performance with a novel constraint programming (CP) approach. We implement an end-to-end framework that enables the evaluation of our methods on instances from the Protein Data Bank (PDB). Our results show that the QUBO approach, using a classical simulated annealing solver, is able to find feasible conformations for problems with up to 6 peptide residues and 34 target protein residues (PDB 3WNE, 5LSO), but has trouble scaling beyond this problem size. In contrast, the CP approach can solve the largest instance in our test set, containing 11 peptide residues and 49 target protein residues (PDB 2F58). We conclude that while QUBO can be used to successfully tackle this problem, its scaling limitations and the strong performance of the CP method suggest that it may not be the best choice.

The Unique Games Conjecture (UGC) has pinned down the approximability of all constraint satisfaction problems (CSPs), showing that a natural semidefinite programming relaxation offers the optimal worst-case approximation ratio for any CSP. This elegant picture, however, does not apply for CSP instances that are perfectly satisfiable, due to the imperfect completeness inherent in the UGC. For the important case when the input CSP instance admits a satisfying assignment, it therefore remains wide open to understand how well it can be approximated. This work is motivated by the pursuit of a better understanding of the inapproximability of perfectly satisfiable instances of CSPs. Our main conceptual contribution is the formulation of a (hypergraph) version of Label Cover which we call label cover. Assuming a conjecture concerning the inapproximability of V label cover on perfectly satisfiable instances, we prove the following implications: * There is an absolute constant c0 such that for k >= 3, given a satisfiable instance of Boolean k-CSP, it is hard to find an assignment satisfying more than c0 k^2/2^k fraction of the constraints. * Given a k-uniform hypergraph, k >= 2, for all epsilon > 0, it is hard to tell if it is q-strongly colorable or has no independent set with an epsilon fraction of vertices, where q = ceiling[k + sqrt(k) - 0.5]. * Given a k-uniform hypergraph, k >= 3, for all epsilon > 0, it is hard to tell if it is (k-1)-rainbow colorable or has no independent set with an epsilon fraction of vertices. We further supplement the above results with a proof that an ``almost Unique'' version of Label Cover can be approximated within a constant factor on satisfiable instances.

Longest‐path routing problems, which can arise in the design of high‐performance printed circuit boards (PCBs), have been proven to be NP‐hard. In this article, we propose a constraint programming (CP) formulation and a mixed integer linear programming (MILP) formulation to gridded longest‐path routing problems; each of which may contain obstacles. After a longest‐path routing problem has been transformed into a CP problem, a CP solver can be used to find optimal solutions. On the other hand, parallel MILP solvers can be used to find optimal solutions after the longest‐path routing problem has been transformed into an MILP problem. Also, suboptimal solutions can be generated in exchange for reduced execution time. The proposed formulation methods can also be used to solve shortest‐path routing problems. Experimental results show that more than 3,700x speed‐up can be achieved by using a parallel MILP solver with 16 threads in solving formulated longest‐path routing problems. The execution time can be further reduced if a computer containing more processer cores is available.

SummaryThis paper describes a learning parallel constraint programming (CP) solver designed for solving CP problems with several instances on massively parallel computing platforms comprising multi‐core parallel machines or Many Integrated Cores. The CP solver proposed in this work is based on aPortfolioparallelization that employs a linear reward inaction learning algorithm in order to obtain the best possible performance for a large set of instances of the same problem. The linear reward inaction algorithm enables the prediction of the number of cores to be assigned to each search strategy based on previous experiments, reducing the computing time required to solve constraint satisfaction and optimization problems. The underlying principle of thePortfolioapproach is to runNsequential search strategies usingNcomputing cores (NToNPortfolio) where each core uses its own strategy in order to perform a search that is different from strategies used by the other cores. The first strategy that finds a solution stops all other strategies. The problem with theNToNPortfolioapproach is that the number of search strategies is very small compared with the current number of computing cores used by the parallel machines. However, using an internal parallelization for each search strategy, it is possible to runNparallel search usingPcomputing cores withP≫N(NToPPortfolio). ThisNToPPortfolioperforms suboptimally for solving different CP problems because many computing resources are wasted. To improve thisPortfoliomodel, an adaptiveNToPPortfoliowas proposed, which tries to privilege the strategy that is most likely to find a solution first in order to give it more computing cores than the other strategies. However, the main problem with the adaptivePortfoliois that it loses all the learned information at the end of each search; it is designed to solve just one CP problem. Furthermore, many computational resources are wasted by bothPortfoliosolvers, the non‐adaptive (NToNandNToP) and the adaptiveNToPPortfolio, when employed in some industrial projects, such as thePAJEROproject, which always solves different instances of the same CP problem. To minimize the amount of wasted resources and to learn the most efficient search strategies, we propose a new learningPortfoliosolver that uses a learning algorithm that configures automatically number of cores to each search. The performance obtained using the differentPortfoliosolvers is compared and illustrated by solving CP problems using as example the Google OR‐Tools solver. Copyright © 2016 John Wiley & Sons, Ltd.

p-Hub location-allocation problem is one of the most interesting subjects in the location theory. Hubs act as switching points to reduce the transportation cost. In this study, two new solution methods, a constraint programming (CP) based model and a hybrid of k-means and genetic algorithm (KGA), are developed to generate exact and approximate solutions, respectively. The proposed CP formulation is more understandable and straightforward in comparison with the MIP model. The experimental results indicate that the CP model uses the memory of the computer (RAM) more efficiently, which enables us to solve the medium size problems. But, in terms of run time, this method cannot be superior to the MIP model. The CP formulation is also extended for the multi allocation p-hub location problem. K-means algorithm, a well-known algorithm for clustering data, is used to generate initial solutions of GA. Furthermore, a new adaptive crossover operator, which is based on the k-means algorithm, is proposed. The experimental results indicate that the KGA algorithm is superior to the GA, regarding time, objective value, and quality of solution measures.

Due to the complexity of nurse rostering problems (NRPs), Constraint Programming (CP) approaches on their own have shown to be ineffective in solving these highly constrained problems. We investigate a two-stage hybrid CP approach on real world benchmark NRPs. In the first stage, a constraint satisfaction model is used to generate weekly rosters consist of high quality shift sequences satisfying a subset of constraints. An iterative forward search is then adapted to extend them to build complete feasible solutions. Variable and value selection heuristics are employed to improve the efficiency. In the second stage, a simple Variable Neighborhood Search is used to quickly improve the solution obtained. The basic idea of the hybrid approach is based on the observations that high quality nurse rosters consist of high quality shift sequences. By decomposing the problems into solvable sub-problems for CP, the search space of the original problems are significantly reduced. The results on benchmark problems demonstrate the efficiency of this hybrid CP approach when compared to the state-of-the-art approaches in the literature.

The Unique Games Conjecture has pinned down the approximability of all constraint satisfaction problems (CSPs), showing that a natural semidefinite programming relaxation offers the optimal worst-case approximation ratio for any CSP. This elegant picture, however, does not apply for CSP instances that are perfectly satisfiable, due to the imperfect completeness inherent in the Unique Games Conjecture. This work is motivated by the pursuit of a better understanding of the approximability of perfectly satisfiable instances of CSPs. We prove that an “almost Unique” version of Label Cover can be approximated within a constant factor on satisfiable instances. Our main conceptual contribution is the formulation of a (hypergraph) version of Label Cover that we call V Label Cover . Assuming a conjecture concerning the inapproximability of V Label Cover on perfectly satisfiable instances, we prove the following implications: • There is an absolute constant c 0 such that for k ≥ 3, given a satisfiable instance of Boolean k -CSP, it is hard to find an assignment satisfying more than c 0 k 2 /2 k fraction of the constraints. • Given a k -uniform hypergraph, k ≥ 2, for all ε > 0, it is hard to tell if it is q -strongly colorable or has no independent set with an ε fraction of vertices, where q =⌈ k +√ k -1/2⌉. • Given a k -uniform hypergraph, k ≥ 3, for all ε > 0, it is hard to tell if it is ( k -1)-rainbow colorable or has no independent set with an ε fraction of vertices.