Prospect Theory in Particle Swarm Optimization for Constraints Nonlinear Optimization Problems

A swarm of individuals or particles work together to achieve a collective goal are called particle swarm optimization (PSO). However, when PSO applied to Constraints nonlinear optimization problems (CNOPs), this technique requires an efficient mechanism to handle the constraints, since it still suffers from the premature convergence problem (local minimum solution), and its inability to find a refinement solution, due to lack of exploitation and exploration capability. So that to deal with the above issues, a novel model of decision-making process has been used in PSO to solve the CNOP. This decision model improves the exploration hence, reduce the risk of premature convergence, by firstly, expanding the search space through generating new alternative solutions and secondly, factoring the risk of violating system constraints during the evolutionary process; while the exploitation has been improved through taking an effective choice (less risky solution) among the generated alternative solutions. The validation of the proposed approach has been performed on several benchmarks CNOPs. The statistical results demonstrate that the new algorithm considerably better than or at least competitive to several evolutionary algorithms reported in the literature.