Design of Semi-chaotic Integration-Based Particle Swarm Optimization Algorithm and Also Solving Travelling Salesman Problem Using It

摘要

Strategy is used to solve nonlinear swarm intelligence optimization problems where social sharing of information helps individual to get benefit from before experience of other companions in search for food. In PSO, solution depends on inertia weight (W), which is an element of acceleration that decides the step size of the solutions. Due to this component, sometimes the global search process may skip the global optima. It suffers from the global minima trap, stagnation and convergence problem. So to avoid this situation and improve the exploration, exploitation and rate of convergence, a new phase is added, named as semi-chaotic integrated PSO (SCIPSO). In SCIPSO, the inertia weight is integrated over a range to enhance the search efficiency. Due to this, solutions are motivated to exploit more desirable in search space. The modified strategy is successfully tested over 25 benchmark functions with other nature-inspired algorithms, and results are best comparatively, which can be further used for the real optimization problem. Along with this, in this paper, the SCIPSO is used to solve one of the widely studied NP-Hard problem named as Travelling Salesman Problem (TSP) to calculate the best minimal cost for maximum of 50 cities whose results are comparatively better than the basic PSO algorithm given under TSPlib.