Research On Optimization Of Yard Crane Scheduling Considering Task Uncertainty In Container Terminals

Port container transportation is an important carrier for international trade in goods.Port container transportation plays an increasingly important role in global trade.As a node connecting sea transportation and land transportation,a container terminal is the hub of the entire port transportation network.In the container terminal,the yard is an important link between the container vessels in the seaside and the domestic customers in the landside.The operation efficiency of the yard crane in the container yard has an important impact on the overall operation performance of the container terminal.At the same time,there are various uncertainties in the practice of yard crane operations.Therefore,this paper considers the uncertainty of the task,and studies the optimal scheduling of the yard crane in the yard.With consideration of the task uncertainty in the yard crane operation,the main works and contributions of this paper are summarized into the following three parts.Rertrieving scheduling optimization under task random arrival time is studied.In the daily operation of the container terminal,the uncertainty of the release times of external trucks is a common phenomenon which has been usually ignored.Moreover,the uncertainty of the arrival for external trucks has a serious impact on the retrieving plan of the yard crane in the container yard.Therefore,considering the uncertainty of the release time of each task and aiming at minimizing the expected total tardiness of all tasks,a two-stage stochastic programming model is established.The first stage is to decide a processing sequence of storage tasks without any information on the release times of the retrieval tasks.In the second stage,the complete information of retrieval tasks is revealed,and the final processing schedule for all the tasks is produced.Furthermore,the uncertainty of task release time is classified,and the sample average approximation algorithm,improved genetic algorithm and rule-based heuristic algorithm are developed to solve the problem.The experimential results show that the designed rule-based heuristic algorithm has stronger convergence and global optimization ability than the sample average approximation algorithm and the improved genetic algorithm.Reshuffling scheduling optimization under task processing time depend on processing scheme is studied.The main incentive for depends on processing scheme is the reshuffling operation.According to the characteristics of the reshuffing operation,it can be divided into two categories.One is called internal reshuffling which means the reshuffle containers are put back to its original stack.Another is called external reshuffling which indicates the reshuffle containers are not put back to its original stack.For the case of internal reshuffling,a mixed integer linear programming model is established with the goal of minimizing the total tardiness of all tasks.A genetic algorithm and a greedy time-based heuristic algorithm are developed.For the case of external reshuffling,a nonlinear model is also established with the goal of minimizing the total tardiness of all tasks.A genetic algorithm and lowest algorithm are also developed to solve the problem.In view of the above results,it has a theoretical guiding role in the design of the reshuffling plan of the yard crane.Delivery scheduling optimization under task alternative depots is studied.Generally,in an automated storage/retrieval system,there is no reshuffling operation.At the same time,the rack and depot locations of each storage task are known beforehand.However,the depot of each retrieval task is unknown and to be determined by many factors.That is to say,the depot location of each retrieval task is uncertain compared to storage tasks.This is also the main decision variable in an automated storage/retrieval system.Therefore,on the basis of considering the release times and due dates of retrieval tasks,a new bi-objective linear mathematical model is established.The two objectives are respectively to minimize the total travel time of the storage/retrieval machine and the total tardiness of all tasks.According to the analysis of the properties of the problem,?-constraint method,non-dominated sorting genetic algorithm and rule-based heuristic algorithm are developed to solve the problem.In conclusion,according to the processing order of a task,the paper analyzes the three task uncertainties.Meanwhile,based on the corresponding properties of the above three optimization problems,the paper establishes respectively corresponding mathematical models that can highlight the characteristic of each problem.At the same time,based on the characteristics of the model and the properties of the problems,the intelligent algorithms are improved and developed.Finally,the paper summarizes the whole contents of the paper,and some future research prospects are proposed.