In this chapter we describe a multi-agent systems (MAS) approach to manufacturing job-shop scheduling that evolves the rules by which schedules are created rather than the schedule itself. The system is tested using a benchmark agent-based scheduling problem and performance results are compared with pure heuristic scheduling heuristics and randomly generated mixed heuristics.

In this paper, we describe the development of a holonic job shop scheduling system that is based on Object-Oriented (O-O) and Multi-Agent Systems (MAS) approaches. In particular, this paper focuses on the holonic architecture (created primarily for system analysis purposes), the multi-agent architecture (created primarily for system design purposes) and the agent platform (Foundation for Intelligent Physical Agents Open Source (FIPA-OS)). On the basis of our experience with this development project, we then reflect on the approach and provide some suggestions for the developers of the holonic systems.

Manufacturing job shop scheduling is a notoriously difficult problem that lends itself to various approaches - from optimal algorithms to suboptimal heuristics. We combined popular heuristic job shop-scheduling approaches with emerging AI techniques to create a dynamic and responsive scheduler. We fashioned our job shop scheduler's architecture around recent holonic manufacturing systems architectures and implemented our system using multiagent systems. Our scheduling approach is based on evolutionary algorithms but differs from common approaches by evolving the scheduler rather than the schedule. A holonic, multiagent systems approach to manufacturing job shop scheduling evolves the schedule creation rules rather than the schedule itself. The authors test their approach using a benchmark agent-based scheduling problem and compare performance results with other heuristic-scheduling approaches.

Merges the latest results obtained by the holonic manufacturing systems (HMS) consortium with the latest developed standards for platform interoperability released by the Foundation for Intelligent Physical Agents (FIPA) to propose a novel e‐business model: the holonic e‐enterprise (HE). The HE extends both the HMS and FIPA models. On one side it extends the holonic manufacturing paradigm with one top level, the inter‐enterprise one. On the other side it extends the multi‐agent system (MAS) paradigm to the hardware (physical machine) level.

Abstract: Merges the latest results obtainedby the holonic manufacturingsystems (HMS) consortium withthe latest developed standards forplatform interoperabilit y releasedby the Foundation for IntelligentPhysical Agents (FIPA) to proposea novel e-business model: theholonic e-enterprise (HE). The HEextends both the HMS and FIPAmodels. On one side it extends theholonic manufacturing paradigmwith one top level, the interenterpriseone. On the other side itextends the multi-agent system(MAS)...

An evolutionary approach to manufacturing job shop scheduling is proposed, implemented through a Multi-Agent System (MAS) exhibiting emergent intelligence. This approach is fundamentally different from traditional approaches by having its focus on the automated scheduler instead of on the scheduling instance, and on flexibility and responsiveness instead of on optimality (at the expense of practicality). The traditional approach to job shop scheduling tries to produce a provably optimal solution to a deterministic problem which then often has to be highly simplified to make it more tractable (i.e. toy problems). Our approach recognizes that the manufacturing environment is inherently concurrent, stochastic, continuous, distributed, and possessing NP-hard complexity. It also recognizes that an optimal solution to a simplified problem is of limited usefulness for a practical implementation of a solution to the original problem. We propose a MAS implementation able to meet the requirements of flexibility, robustness, and scalability, as well as to support the emergent nature of the intelligence in this learning system.