A Novel Particle Swarm Optimization Algorithm for Solving Transportation Problem

The transportation problem (TP) is well known as a basic network problem for it could be extensively applied in many fields. The linear transportation problem (LTP), which is the core and basic model of TP, can be extended to other TP with higher complexity. In the present paper, a new particle swarm optimization algorithm (PSO-TP) whose special structure and operators are different from the classical PSO is given for the solution to LTP. A new position updating rule and a negative repair operator of PSO-TP can help to meet the constraints of LTP, which consequently saves much computational cost to find the feasible solution. Moreover, a mutation operator is added to enable PSO-TP not to finish searching prematurely. Numerical experiments show the effectiveness and efficiency of the proposed algorithm, through the comparison with Vignaux and Michalewicz´s genetic algorithm (GA) and the performance in solving open problems