Portfolio approaches in constraint programming

Abstract

Recent research has shown that the performance of a single, arbitrarily efficient algorithm can be significantly outperformed by using a portfolio of -possibly onaverage slower-algorithms. Within the Constraint Programming (CP) context, a portfolio solver can be seen as a particular constraint solver that exploits the synergy between the constituent solvers of its portfolio for predicting which is (or which are) the best solver(s) to run for solving a new, unseen instance.In this thesis we examine the benefits of portfolio solvers in CP. Despite portfolio approaches have been extensively studied for Boolean Satisfiability (SAT) problems, in the more general CP field these techniques have been only marginally studied and used. We conducted this work through the investigation, the analysis and the construction of several portfolio approaches for solving both satisfaction and optimization problems. We focused in particular on sequential approaches, i.e., single-threaded portfolio solvers always running on the same core.We started from a first empirical evaluation on portfolio approaches for solving Constraint Satisfaction Problems (CSPs), and then we improved on it by introducing new data, solvers, features, algorithms, and tools. Afterwards, we addressed the more general Constraint Optimization Problems (COPs) by implementing and testing a number of models for dealing with COP portfolio solvers. Finally, we have come full circle by developing sunny-cp: a sequential CP portfolio solver that turned out to be competitive also in the MiniZinc Challenge, the reference competition for CP solvers.iii iv

A Novel Approach to String Constraint Solving

“…Analogously, G +S is available on-line [119] but no longer developed. Conversely, the G-Strings solver [9] implementing the dashed string approach is actively maintained. Both G +S and G-Strings are extensions of Gecode [62], a well-established CP solver over finite domains.…”

Equitable optimized airspace sectorization based on constraint programming and OWA aggregation

“…Recent research in artificial intelligence areas, such as CP (Rossi et al, 2006), Boolean satisfiability (SAT) testing (Xu et al, 2007) and integer linear programming (ILP) (Wolsey, 1998), has shown that the performance of a sequential solver can be significantly improved by carefully exploiting a parallel solving architecture. In the literature, we can distinguish two mains approaches to exploit parallelism: divide-and-conquer and cooperative portfolio strategy (Amadini, 2015). In the former, the problem is divided into several sub-problems, each problem is solved separately and the result is later combined to form the solutions of the original problem, whereas the parallel portfolio strategy exploits the complementarily between different sequential solving strategies to let them compete and cooperate on the same problem.…”

A Constraint-Based Model for Fast Post-Disaster Emergency Vehicle Routing

“…We did not find in the literature suitable and extensive benchmarks of disaster scenarios that we could use to evaluate and compare the performances of our approach. For this reason, in order to evaluate our algorithm we extended the methodology used in [9]. In particular, we built random generated scenarios obtained by varying the number of hospitals in the set {1, 2, 4}, the number of ambulances in {4, 8, 16, 32, 64}, and the number of victims in {8, 16, 32, 64, 128, 256, 512}.…”