Development of multi-objective optimization methods for integrated scheduling of handling equipment (AGVs, QCs, SP-AS/RS) in automated container terminals

Increasing demand for containerization and large containerships in service compel container terminals to improve their performance. In automated container terminals (ACTs), containers are unloaded from (loaded to) the vessels, by using quay cranes (QCs). Commonly, containers are stacked in storage yards by using automated stacking cranes (ASCs) before they delivered to final customers, or to the arriving vessels. Moreover, automated guided vehicles (AGVs) are used to connect the QCs to the storage yard. In order to increase land utilization and lower times for the storage and retrieval operations, a new storage system called split-platform automated storage/retrieval system (SP-AS/RS) has been introduced. In this system, containers are stored in racks, and two platforms are used to transfer them horizontally and vertically. Previous researches showed that integrated scheduling of handling equipment helps ACTs to improve their performance. However, there is no integrated scheduling method for QCs, AGVs, and SP-AS/RS in previous researches. In this research, the integrated scheduling is described as a multi-objective problem. A mixed integer-programming (MIP) model was developed to optimize the integrated scheduling of handling equipment with the objective function of minimizing total delays in tasks of QCs, in addition to total travel time of AGVs and platforms of the SP-AS/RS. The comparison of results for this method and an ACT in which ASCs are in use indicates that delays in tasks of QCs are reduced by 8.6%, on average. Moreover, the value of the objective function obtained by the proposed integrated scheduling method decreased 58%, on average, compared against nonintegrated scheduling method. The integrated scheduling has been proved as a nondeterministic polynomial-time hard (NP-hard) problem, which means there is no systematic approach to obtain the optimal solution especially for relatively large cases, in reasonable computation time. Therefore, two meta-heuristic algorithms, namely genetic algorithm (GA) and simulated annealing (SA) algorithm, were developed to optimize the integrated scheduling of handling equipment. The metaheuristic algorithms determine sequence of loading/unloading tasks; and a heuristic rule assigns the AGVs to the tasks. Results indicate that “earliest available vehicle” is the best heuristic rule for the integrated scheduling method. Moreover, it is shown that on average, the best objective values obtained by the GA and SA algorithm, are only 6.4% and 3.7% worse than the optimal ones found by the MIP model, respectively; demonstrating that both algorithms are able to achieve near optimal solutions. However, the GA outperforms the SA algorithm in medium and large size test cases. Sensitivity analysis shows that value of the objective function decreases as the number of available AGVs increases. Furthermore, it is revealed that to decrease delays in tasks of the QCs, the AGVs have to travel longer routes.

Preview PDF FK 2012 68R.pdf Download (1MB) | Preview

Actions (login required)

View Item