船席指派與橋式起重機指派是貨櫃碼頭的兩大船邊作業問題。有效解決此兩大問題,可提升貨櫃碼頭作業之效率。過去之研究,基因演算法是解決貨櫃碼頭船邊作業問題的主要方法。近年來,粒子群演算法(Particle Swarm Optimization, PSO)崛起,並逐漸被應用於解決組合最佳化問題。但將其應用在解決貨櫃碼頭船邊作業問題之研究則仍罕見。因此,本研究提出一個混合式粒子群演算法 (Hybrid PSO, HPSO)來同時解決「動態及離散型」船席指派與「動態式」橋式起重機指派問題。在結合動態排序及離散事件技術後,HPSO並且可進行「變動式」橋式起重機指派。在與傳統之基因演算法(Genetic Algorithm, GA)比較後,此HPSO顯示了較佳之規劃結果。本研究中之GA係採用兩點交配及交換突變運算,並且只能進行「固定式」橋式起重機指派。
Berth allocation problem (BAP) and quay crane assignment problem (QCAP) are two essential seaside problems in a container terminal (CT). They can impact the performance of a CT significantly due to the uses berth and quay crane, two scarce resources in a CT. It is noted that genetic algorithm (GA) have been playing the main role in dealing with the two problems. However, particle swarm optimization (PSO) has been considered as a good competitor to GA in solving combinational optimization problems (COPs). But it has never been used to deal with the two problems in terms of variable-in-time QC assignment. This has prompted us to propose a new hybrid PSO (HPSO) to deal with the “dynamic” and “discrete” BAP (DDBAP) and “dynamic” QCAP (DQCAP) at the same time. The HPSO hybridizes a PSO with a simulated-based heuristic and, together with the techniques of dynamic rank order values (DROVs) and discrete events, it can perform variable-in-time QC assignment. To investigate its effectiveness, the HPSO has been compared to a GA that employs two-point crossover (TPX) and swap mutation (SWM) operations. The results of the experiments show that the HPSO outperforms the GA in terms of fitness value (FV).