Modified PSO Based on BVA Techniques

Knapsack is a combinatorial problem used to solve the multiple constrain problems by considering weight and capacity, various algorithm exist to solve the knapsack problem such as dynamic program, hybrid approach, branch and bound. Still multi-constraint knapsack problem remains the major challenge. Particle swarm optimization is an optimization problem which optimizes a solution efficient by considering only few parameters there by solving multi-constraint problem efficiently. In this paper we have proposed the modified PSO algorithm which solves the knapsack based on multi-constraint (weight and capacity), proposed algorithm is based on Boundary Value Analysis technique and is compared with the standard benchmark dataset of knapsack problem, the computation time, minimum deviation, average deviation, standard deviation are evaluated. The performance of normal knapsack problem and the proposed algorithm is measured. The proposed algorithm performs better compare to normal PSO knapsack problem. In future we plan to extend our work for other combinatorial optimization problem.