Article ID: | iaor20081906 |
Country: | Netherlands |
Volume: | 42 |
Issue: | 4 |
Start Page Number: | 2186 |
End Page Number: | 2196 |
Publication Date: | Jan 2007 |
Journal: | Decision Support Systems |
Authors: | Lim Andrew, Wang Fan |
Keywords: | stochastic processes |
In this paper, the optimization of the Berth Allocation Problem (BAP) is transformed into a multiple stage decision making procedure and a new multiple stage search method, namely stochastic beam search algorithm, is proposed to solve it. New techniques such as an improved beam search scheme, a two-phase node goodness estimation, and a stochastic node selection criteria are proposed. Real-life information provided by Singapore Port was collected as our test data. Experimental results show that the proposed stochastic beam search is more accurate and efficient than both the state-of-the-art meta-heuristic and the traditional determinist beam search.