A hybrid algorithm using particle swarm optimization for solving transportation problem

Particle swarm optimization (PSO) is a well-known population-based stochastic optimization algorithm intended by collective and communicative behavior of bird flocks looking for food. Being a very powerful tool for obtaining the global optimal solution, PSO has experienced a multitude of enhancements during the last three decades. The algorithm has been modified, hybridized and extended by various authors in terms of structural variations, parameters selection and tuning, convergence analysis and meta-heuristics. In this article, hybridized PSO has been proposed to solve balanced transportation problem, a discrete optimization problem, of any number of decision variables converging to the global optima. Two additional modules have been embedded within the PSO, in order to repair the negative and/or fractional values of the decision variables, and tested with variants of parameters present therein. The proposed algorithm generates an optimal solution even without considering the rigid conditions of the traditional techniques. The paper compares the performance of different variants of inertia weight, acceleration coefficients and also the population size with respect to the convergence to the optimal solution. The performance of the proposed algorithm is statistically validated using the pairedttest.