Robust Optimization Of Port Berth Allocation In Uncertain Environment

Shipping is a vital component of global trade and economic development.It is the most important mode of transportation in the logistics industry.Shipping plays a vital role in trade,energy,tourism and so on.The container terminal plays an indispensable role in sea transportation.The container terminal is specially used for loading and unloading container goods,and its role and importance are very great.In recent years,with the increasing proportion of sea transportation,higher requirements are put forward for container terminals.In this context,how to reasonably arrange the operation of container terminals is particularly important.In order to solve the operational optimization problem of container terminals,it is necessary to consider the allocation of berths and quayside Bridges.Since the resources of berth and quayside Bridges in container terminals are always limited,reasonable scheduling of berth and quayside Bridges can improve the operation efficiency of container terminals and reduce economic and time costs.Therefore,berths and quayside bridge scheduling is the core of container terminal operation.Based on domestic and foreign studies on improving berth allocation and quay bridge allocation,this paper establishes a biobjective robust optimization model with the objective of minimizing berth waiting time and optimal position deviation,considering the uncertainty of ship arrival time.At the same time,considering berth allocation and quayside bridge allocation for joint scheduling,and establish a multi-objective berth berth joint scheduling robust optimization model under uncertain conditions to minimize ship waiting time,ship departure delay time and ship optimal berth deviation.Different from random programming,robust optimization does not need to know the specific distribution of uncertain parameters,but represents the uncertainty of parameters with multi-scenario sets.Based on genetic algorithm,this paper introduces the local search idea of simulated annealing algorithm to improve the traditional genetic algorithm and improve the local mining ability of traditional genetic algorithm.Port ship arrival time and deviation time were collected as sample data.The K-S test method was used to conduct non-parametric test on ship arrival time distribution and ship arrival time deviation distribution.It was concluded that ship arrival time obeyed exponential distribution and arrival time deviation obeyed normal distribution,and then the required data under multiple scenarios were generated by exponential distribution and normal distribution for experiments.The effectiveness of the model and the algorithm is verified by numerical experiments.It is found that the robust optimization model constructed by the variance and expectation of the objective function of the multi-scenario set is more stable to the fluctuation of uncertain parameters than that constructed by the expectation of the objective function of the multi-scenario set,which improves the response ability of the model to uncertain parameters.Meanwhile,compared with the traditional genetic algorithm,the convergence speed of the designed hybrid genetic algorithm is faster,and the quality of the solution obtained by the adaptive genetic algorithm is also improved to a certain extent with the increase of the problem scale.