Using fast matrix multiplication in bio-inspired computation for complex optimization problems

Population-based search heuristics such as evolutionary algorithms or ant colony optimization have been widely used to tackle complex problems in combinatorial optimization. In many cases these problems involve the optimization of an objective function subject to a set of constraints which is very large. In this paper, we examine how population-based search heuristics can be sped up by making use of fast matrix multiplication algorithms. First, we point out that this approach is applicable to the wide class of problems which can be expressed as an Integer Linear Program (ILP). Later on, we investigate the speedup that can be gained by the proposed approach in our experimental studies for the multidimensional knapsack problem.