Optimization On Relocation Problems Of Outbound Containers In Container Terminals

With the rapid growth of containers throughput, the scarcity property of terminal yards is becoming increasingly prominent. Therefore, how to improve the efficiency of terminal operations, accelerate the transfer of containers in the yard, and maximize the use of existing space resources and equipment resources, have become an urgent problem for terminal managers. Aimed at solving the relocation problem, the key factor for the operating efficiency of the yard, a large number of related literatures are reviewed and summarized. The solutions on the relocation problem are still not satisfying. Relatively little work has been done on dynamic characteristics. The main work in this dissertation about the above problems is as follows:(1) Considering the uncertainties of export containers, to minimize the follow-up relocations, an integer programming model is established. The container static retrieval problem is studied under the circumstance that the stacking state is given without external arrival containers. Based on the static solving methods, the dynamic problem is discussed. Combined the accurate algorithm with heuristic rules, the existing algorithms are improved and can better guide the actual operation.(2) Based on the actual situation and realistic constraints, a loading sequence optimization model is formulated to minimize the number of relocations. A two-phase hybrid optimization algorithm is presented, and to avoid combinatorial explosion of states, heuristic rules are embed into the dynamic programming algorithm. Computational experiments showed the efficiency and superiority of the proposed model and algorithm, compared with the actual operation rules and existing researches.(3) A two-stage hybrid algorithm is proposed to solve the pre-marshalling problem. In the first stage, the final layout state space is compressed by a neighborhood search algorithm heuristic rules, and in the second stage, the length of pre-marshalling sequence is shortened by an integer programming. Two stages execute alternately to get optimal pre-marshalling sequence as soon as possible. Experimental results demonstrated the effectiveness and practicability of the hybrid algorithm.(4) The single-spreader-single-crane, the multi-spreader-single-crane and the single-spreader-multi-crane retrieval problems are discussed. To minimize the movements of containers and to shorten the makespan of cranes, heuristic algorithms are proposed to solve the three problems respectively. Numerical experiments showed the effectivity and superiority of the algorithms compared with existing researches.