Binary Satin Bowerbird Optimizer for the Set Covering Problem

Abstract

The set covering problem (SCP) is one of the most studied NP-hard problems in the literature. To solve the SCP efficiently, this study considers a recently proposed bio-inspired meta-heuristic algorithm, called satin bowerbird optimizer (SBO). Since the SBO was first introduced for the global optimization problem, it works on a continuous solution space. To adapt the algorithm to the SCP, this study introduces a binary version of the SBO (BSBO). The BSBO simply converts real value coded solution vector of the SBO to binary coded solution vector by applying a binarization procedure. In addition to binarization procedures, a solution improving operator is employed in the BSBO to transform infeasible solutions into feasible solutions and remove redundant columns to reduce solution cost. The performance of the proposed BSBO is tested on a well-known benchmark problem set consists of 65 instances. With regards to the best-known solutions of the instances, efficient results are obtained by the proposed BSBO by finding near-optimal solutions. Furthermore, standard deviations of the runs demonstrate the robustness of the algorithm. As a consequence, it should be noted that the proposed solution approach is capable of finding efficient results for the SCP.