Scheduling optimization in restricted channels based on the agent technology and Bayesian Network

In ports and inland waterway, there are a considerable proportion restricted channels, in which vessels must navigate in a set route and time. The transit capacities of inland waterways and ports are limited by those restricted channels significantly, which can only be scheduled by experienced supervisors of local harbor and maritime administration manually. In fact, the scheduling optimization requires much experience and knowledge. Hence, not mangy supervisors are capable of optimizing those channels satisfactorily. This paper proposed an optimization scheduling approach based on the Agent technology and Bayesian Network. At the very beginning, three dimensional model of a specified channel is developed. Subsequently, the characteristics of vessel behaviors are extracted from AIS data with the help of Bayesian Network. On that basis, the models of vessel behaviors are built by the Agent technology. Different alternatives of scheduling can be simulated. Eventually, the performance of different scheduling might be obtained in the simulation of environment. This approach is based on data mining, Agent simulation and Bayesian Network; it does not need the manual experience any longer, which is capable of choosing the efficient scheduling plan in different scenarios. This approach improves the transit capacity of inland waterways and ports.