The paper describes functionality of Magenta Multi-Agent Logistics i-Scheduler Engine presented on AAMAS 2006 onferences and gives examples of its application in business domain.

The i-Scheduler Engine was designed to be scalable without risk of combinatorial explosion, in order to handle large transportation networks as a whole. The multi-agent architecture combined with semantic network allows very granular approach for every business entity of transportation network (client, order, cargo, truck, driver, etc) and balancing of their conflicting interests. The i-Scheduler considers individual constraints and, interestingly, specific preferences of customers, drivers, trucks, cargoes, etc. This results in a unique ability to combine inbound and outbound deliveries, different fleets or private networks, driving more value from finding effective backhauls and consolidations.

The paper covers the history of development, architecture and current functionality of the engine and provides a set of case studies in different transportation networks, which outline the most serious challenges Magenta overcame in each case.