Discrete Chicken Swarm Optimization for the Quadratic Assignment Problem

The quadratic assignment problem (QAP) is a well-known combinatorial optimization problem, which could be applied to different applications. The main objective of this paper is to present the first discretization of the chicken swarm optimization algorithm (CSO) to solve quadratic assignment problem without using a local search, the adaptation of CSO in discrete case is based on redefining operations and operators of the original version. As known, the CSO is a stochastic method inspired from the behavior of chickens in swarm while searching for food. The experiments are performed on a set of 56 benchmark QAPLIB instances. To prove the choice of the adequate parameters, a study is conduct on CSO using simulations on certain instances. The discussion of different tests obtains competitive results compared with the known metaheuristic of Genetic algorithm based on SCX. The results demonstrate effectiveness of the proposed CSO-QAP to solve the quadratic assignment problem in term of time and quality of solutions. The proposed adaptation can be further applied by using a local search strategy such as 2-opt in order to solve the same problem or another NP-hard combinatorial problem