Decomposition Based on Decision Diagrams

Abstract

Always cite the published version, so the author(s) will receive recognition through services that track citation counts, e.g. Scopus. If you need to cite the page number of the author manuscript from TSpace because you cannot access the published version, then cite the TSpace version in addition to the published version using the permanent URI (handle) found on the record page.This article was made openly accessible by U of T Faculty. Please tell us how this access benefits you. Your story matters. Abstract. In recent years, decision diagrams (DDs) have proven useful for solving a variety of optimization problems, often closing long-standing instances from classical benchmarks. This success is primarily driven by a DDs ability to capture structure. This paper exploits this characteristic and proposes a novel solution method which decomposes a problem into highly-structured portions, where the solution set of each portion can be compactly represented using a DD. This technique is applied to a special case of the independent set problem and to unconstrained binary quadratic programming. Preliminary computational results suggest that the proposed decomposition approach can improve upon both standard integer programming models and a single DD approach. Decomposition Based on Decision Diagrams

What's Hot at CPAIOR (Extended Abstract)

“…(Chu, Stuckey, and Gange 2016) use a constraint solver to solve sub-problems and then use a Lagrangian decomposition to compute a better bound on the objective function. (Bergman and Cire 2016) decompose a problem into decision diagrams that can be efficiently handled.…”

FastDOG: Fast Discrete Optimization on GPU

“…The works [4,5,38] similarly consider decompositions into multiple BDDs and solve the resulting problem with general purpose ILP solvers. The work [6] investigates optimization of Lagrange decompositions with multi-valued decision diagrams with subgradient methods.…”

A Benders Approach to the Minimum Chordal Completion Problem

“…Relaxed caveman graphs (Judd et al 2011) represent typical social networks, where small pockets of individuals are tightly connected and have sporadic connections to other groups. This family of instances has been employed previously in the evaluation of algorithms for combinatorial optimization problems (Bergman and Cire 2016). Each instance is generated randomly based on three parameters, α, β ∈ Z + and γ ∈ (0, 1).…”