An application oriented multi-agent based approach to dynamic load/truck planning

• A novel multi-agent system for dynamic load and truck planning.
• A realistic simulation with real data from a major transportation company.
• A practical negotiation based model for real life implementations.
• Detailed explanations and numerical examples are provided to enable a better understanding of the model.

Truck operations decisions for transportation logistics pose challenges especially when loads are less-than-truckload (LTL). Within a dynamic business environment load planners should consider effective utilization of resources and profitability of their operations. Multi-agent based system provides effective mechanisms for the management of dynamic operations in transportation. The algorithms for transportation domain that are available in the literature are generally focusing on generation of effective solutions for planning/scheduling problems without considering real transportation systems dynamics. Multi-agent based design of the load/truck planning problems is supposed to be helpful for integration of algorithms with real-time logistics controlling systems. The cooperative structure of the multi-agent based approach is motivated by real-world third party logistics (3PL) company operations. Negotiation mechanism among the agents is used to handle the dynamic events. The proposed approach is tested via simulation by using LTL data from a 3PL logistics company. The approach generates feasible and profitable decisions under dynamic circumstances by using negotiation/bidding mechanisms. Proposed approach is implemented by using JACK™, an agent development framework. A multi-agent based dynamic load/truck control system (MABDLCS) is also developed along with this approach. MABDLCS could be used for both testing some transportation scenarios and for real time vehicle/load control purposes. The solutions obtained by using the proposed approach demonstrated that MAS is contributing on problem solution quality while generating real-time schedules.