A simulation based hybrid algorithm for yard crane dispatching in container terminals

Abstract

The problem of yard crane dispatching in container terminals is addressed in this paper. We proposed two new hybrid algorithms which combine the advantages of A* heuristic search and Recursive Backtracking with prioritized search order to accelerate the solution process. The algorithms proposed use real time data-driven simulation to accurately predict the time taken by the yard crane in performing its operations and this helps in getting an optimal dispatching sequence that can be followed by the yard crane. Experiments carried out show that the proposed algorithms consistently perform very well over all tested cases. The best performing algorithm is able to find the optimal solution over 2.4x10 possible dispatching sequences in about 0.3 to 0.4 seconds under heavy workload. The characteristics of memory-saving and interruptibility enable the algorithm to be easily integrated into a complete yard crane management system in real world applications. In such real time yard crane management system, our proposed algorithms can be used as an effective and efficient tool to support complex and intelligent higher level planning in addition to managing the yard crane operations in its appointed zone.