This paper addresses the berth allocation problem in automobile rollon/rolloff(Ro-Ro)ter-minals,considering conflicts in the flow of automobiles.Initially,the study defines conflicts in the forefront vehicle flow at Ro-Ro terminals,introducing a conflict penalty time coefficient to quantify the associated penalty time.We then proposed two mixed-integer programming models:one allow-ing automobile flow conflicts and the other prohibiting them.Both models aim to minimize the total working time of the ships.Subsequently,we design a genetic algorithm(GA)incorporating a multi-layer correspondence encoding for Ro-Ro ships and automobile flows.Additionally,we implement a decoding algorithm based on a rule strategy.A total of four sets of test cases of varying scales are gen-erated for numerical experiments,and the parameters of the GA are determined based on experimen-tal results.In small-scale cases,the optimal solution of genetic algorithm is within 1%of the exact CPLEX solution.In large-scale cases,the GA outperforms the strategy-based solution derived from empirical scheduling,thus validating the correctness of the model and the effectiveness of the GA.Fi-nally,this study conducts sensitivity analysis of conflict penalty coefficients under two conflict strate-gies.The results indicate that when the coefficient is below 0.5,a noticeable deviation is observed in the objective function values of the two strategies.Hence,when formulating berth allocation plans for automobile Ro-Ro terminals,the impact of conflicts between automobile flows should be consid-ered,particularly when the conflict penalty coefficient is low.
维普
万方数据