Container terminal is the transfer hinge of land-sea intermodal transportation. Container terminal logistics system (CTLS) is the central handling platform of container logistics in logistics network as the computer system is the core processing node to information flow in internet. Quay side is its “central processing unit (CPU)” of handling platform and storage yard is its “main memory.” CTLS is a highly complex system which includes parallel, negotiation, and competition relations etc. in operation so that this paper introduces agent based computing (ABC) and knowledge discovery in database (KDD) to model CTLS.

Logistics, information flow, and decision flow are the kernel of transaction and operation of container terminals which are developing into professional third party logistics (TPL) and seized of enormous business data which are accumulated by the inner information infrastructure and the public electronic commerce (EC) platform. Through the analysis to the handling operation and container logistics of CTLS, we refer to the design thinking of computer organization and architecture and introduce the resource scheduling thinking of the operation system, and draw an analogy between CTLS and parallel computer architecture based on Harvard architecture, map and carve up the production, and scheduling system of CTLS into fourteen kinds of agents: berth allocation agent, quay crane allocation agent, storage yard management agent, yard crane allocation agent, container truck dispatch agent, tugboat dispatch agent, gate house management agent, berth agent, quay crane agent, yard agent, yard crane agent, container truck agent, tugboat agent, and gate house agent according to the entities and their mutual relations. We bring forward the modeling framework of CTLS based on multi-agent and KDD which provides the engine to drive the model accordingly.

This modeling thinking and framework, which serves the turn to the queuing theory inherently and possesses the top-ranking flexibility and robustness to the computational intelligence (CI), can help to set down the production scheduling plan and support the important decision-making of the port. A simulation to the quay side is implemented by the AnyLogic 6.2 and SQL Server 2005 and berth allocation based on practical data is given emphasis to discuss. The simulation result precedes the quondam research and actual data to 5.14% and 26.82% through the dynamic priority scheduling. In addition, a series of similar simulations are executed to support the decision on the proportion between berth and quay crane. All verify the feasibility and creditability of the above idea and the effective scheduling strategy in the framework will be further discussed and studied when the container terminal is provided with the multifarious characteristics in the future.