Self-adaptive Systems: Facilitating the Use of Combinatorial Problem Solvers

Abstract

New methods in Combinatorial Problem Solving can solve larger problems in different domains. They also became more complex, which means that they are hard to use and fine-tuning to the peculiarities of a given problem, limiting its use to a small set of experts, and instead black-box solvers with automated search procedure are needed for its broad applicability. Autonomous Search Systems represent a new research field defined to precisely address the above challenge. The main goal of this paper is to review recent works on this kind of Self-adaptive Systems from the standpoint of the actual requirement for solvers.

Nature-inspired Parameter Controllers for ACO-based Reactive Search

“…As combinatorial optimization solvers, Ant Colony Optimization (ACO) algorithms are accorded much work to improve their efficiency and usability. Parameter adaptation and reactive search are autonomous search techniques for improving the behavior of the algorithms by integrating control in the solving process (Broderick et al, 2014). The control takes either a machine learning fashion, as in reactive search, or the self-tuning fashion, as in parameter adaptation.…”