A simplified binary artificial fish swarm algorithm for 0–1 quadratic knapsack problems

Abstract

This paper proposes a simplified binary version of the artificial fish swarm algorithm (S-bAFSA) for solving 0–1 knapsack problems. This is a combinatorial optimization problem, which arises in many fields of optimization. In S-bAFSA, trial points are created by using crossover and mutation. In order to make the points feasible, a random heuristic drop item procedure is used. The heuristic add item is also implemented to improve the quality of the solutions, and a cyclic reinitialization of the population is carried out to avoid convergence to non-optimal solutions. To enhance the accuracy of the solution, a local search is applied on a predefined number of points. The method is tested on a set of benchmark 0–1 knapsack problems.