Multi-population Hybrid Shuffled Frog Leaping Algorithm And Its Application To Path Planning Of Gantry Cranes At Container Yards

The rapid development of global economy makes trade transportation rise sharply.Ship transport plays a very important role in trade transportation.Particularly,container transportation is the most important one among shipping modes.In order to improve their economic benefit,container terminals must improve their work efficiency.Among those,operating efficiency of gantry cranes at container yards is one of the key factors that affect the whole efficiency of container terminals.To solve the path planning problems of gantry cranes perfectly can effectively improve the working efficiency and whole benefit of container terminals,so as to enhance the competitiveness of ports.Swarm intelligent optimization algorithms are adopted to solve the path planning problems of gantry cranes at container yards in this paper.Firstly,according to the given pick-up order of gantry cranes,mathematical models aimed at minimizing the working path of a single gantry crane and multiple gantry cranes are constructed respectively while the yard deposit characteristics are known.Then,an efficient algorithm is developed to solve this discrete combinatorial optimal problem.Based on the relatively novel Shuffled Frog Leaping Algorithm and to overcome some defects of it,a Multi-population Hybrid Shuffled Frog Leaping Algorithm(MHSFLA)is put forward.The fundamental idea of the proposed algorithm is dividing the whole population into three subpopulations with the same size.One subpopulation focuses on learning from the global optimum to accelerate the convergence speed;another subpopulation focuses on local search near superior individual;and the last one focuses on global searching,maintaining the population diversity and prevent premature.Regular information exchange is conducted among the three subpopulations to make full use of complementary advantages of each one and improve the whole performance of proposed algorithm.In addition,the introduction of the crossover and mutation operators of Genetic Algorithm makes proposed algorithm suitable for the discrete combination optimization problems,such as these path planning problems.To be hybridized with Simulated Annealing algorithm can improve the local search ability around the optimal individuals,accelerate convergence rate and prevent the proposed algorithm from premature.The proposed algorithm is applied to solve the classical function optimization and traveling salesman problem whose optimal solutions are known to verify its performance.And the optimal results illustrate its feasibility and effectiveness.Based on above work,taking the path planning problems of gantry cranes at container yards as engineering background,the specific encoding scheme and the crossover and mutation strategies are designed for applying the proposed algorithm to the mathematical model presented above.Two path planning examples are solved and tested by proposed algorithm under the circumstances of single gantry crane and two gantry cranes.And the simulation results are compared and analyzed.It shows that the proposed algorithm is effective for the path planning problems of gantry cranes at container yards.It can give superior path planning schemes which are satisfied for engineering application.This work can provide references for the actual operation of gantry cranes at container yards,and contribute to higher working efficiency and economic benefits for container terminals.The research has certain theoretical significance and practical application value.