Decide : agents controlling a BHS of an airport hub

It is an ingrown fsar to most travelers, frequent or not, to lose their baggage while flying from one destination to another, but if you take a look behind the scene you will experience a system with a complexity unlike most others. The baggage handling system (BHS) of an airport is just one link in a Song chain of processes your bag is going through. The BHS stretches from check-in or unloading of a plane on to the BHS (connecting flights), to it is collected at the gate and transferred to your flight. The conventional control software use a strategy primarily based on a shortest path algorithm, not taking into account dynamical changes or utilizing less packed areas of the system to increase capacity. We changed that perspective towards a ecentralized multi-agent based solution by developing strongty collaborating agents, to replace the original control software. The hierarchy of agents horizontally spans the system through local agents capable of altering the flow of totes, and vertica!ly through mediator agents, which have a global perspective andsupport decisions ofthe locoi autonomous agents. The local agents monitors their own local neighborhood for rising queues and other delaying factors, which aggregate to a status of the node, influencing the decision, of other agents trying to loute totes through this node. Interactions are handled as a combination of contract-nets, queries, and subscriber interaction schemes, which generalized the agent interfaces from the specific application. An agent-based approach not only improves robustness of the system, and utilize the entire BHS in n more convenient and dynamical way, it also include strategies for maximizing capacify of the system, which follows a work in-progress against capadty curve (WIPAC). In this paper we present the agent-based design of the control-software, and elaborate on the mutual colloboration compliant with the FIPA interaction schemes. We present ontologies used for sharing information and services among the agents as well as the results of our work for a major airport hub in Asia, and compare it against the traditional centralized strategies. We complete the paper by drawing same condusions and present ideas for future work.