| Summary: | The work presented in this thesis is motivated by the 'hard' nature of the job shop scheduling task due to the 'intrinsic' and 'extrinsic' complexity of realistic problems. A scheduling framework, combining Artificial Intelligence and Operations Research techniques and implemented in a Distributed Problem Solver environment suitable for parallel implementation, is described. The adopted approach views the system as an <i>Organisation</i>. Agents are assigned different roles and functions depending on their position within the structure of the <i>Organisation</i>. In this Organisation, agents of the same level state their interests independently of each other and therefore <i>Conflict</i> is likely to occur. A major thesis of the research reported here is that not only is it important to deal with <i>conflict</i> but also that <i>conflict</i> as a consequence of the scheduling process should be exploited as a way of integrating different scheduling perspectives, as a way of allowing agents to express their own interests independently of each other and, thus, guaranteeing <i>pluralism. Pluralism</i> is also ensured by providing agents with both empirical knowledge (heuristics, dispatch rules) and theoretical knowledge (optimal algorithms) and by explicitly allowing the coexistence of a <i>job based perspective</i>, a <i>resource based perspective</i> and an <i>operation based perspective</i> enabling so called <i>opportunistic</i> and <i>micro-opportunistic</i> scheduling. In order to achieve <i>Global Coherence</i> in this conflicting distributed environment, agents are provided with mechanisms to make them aware of the structural and intrinsic features of the (sub)problems that they have to solve and the interaction of their (sub)problems, without relying on communication with each other, and with tools to analyse, evaluate and solve the <i>conflicts. Structural Awareness</i> is a major concept introduced and developed in the research reported in this thesis.
|
|---|
