Scheduling, re-scheduling and communication in the multi-agent extended enterprise environment

Abstract

In this article we describe a multi-agent dynamic scheduling environment where autonomous agents represent enterprises and manage the capacity of individual macro-resources in a production-distribution context. The agents are linked by client-supplier relationships and inter-agent communication must take place. The model of the environment, the appropriate agent interaction protocol and a cooperative scheduling approach, emphasizing a temporal scheduling perspective of scheduling problems, are described. The scheduling approach is based on a coordination mechanism supported by the interchange of certain temporal information among pairs of client-supplier agents involved. This information allows the agents to locally perceive hard global temporal constraints and recognize non over-constrained problems and, in this case, rule out non temporally-feasible solutions and establish an initial solution. The same kind of information is then used to guide re-scheduling to repair the initial solution and converge to a final one. © Springer-Verlag Berlin Heidelberg 2001.