Abstract
Production planning and resource allocation is a complex industrial decision-making problem. Sophisticated computational model of a manufacturing domain may support this decision making by simulation of multiple variants of alternative plans and thus help identifying the most suitable one (according to defined conditions). Multi-agent system is an example of such a computational model as it can naturally represent the hierarchical and distributed structure of the manufacturing enterprise that is modelled. This item presents and discusses ProPlanT (Marík, Pechoucek, Štepánková, & Lažanský, 2000), a specific multi-agent technology/methodology for production planning and scheduling in the manufacturing domain. This methodology resulted in a framework for decision making support which was successfully applied in several pioneer applications.
Key Terms in this Chapter
ProPlanT: Multi-agent technology/methodology for decision-making support in the manufacturing domain. It supports production planning, scheduling, and simulation on the intra-enterprise and extra-enterprise level.
Intra-Enterprise (IE) Agent: An agent operating on the intra-enterprise level of the system.
Extra-Enterprise Level: The part of the system operating outside of boundaries of a company. On this level, the cooperation with the customers/suppliers, supply chain planning and optimization, and simulation is taken into account.
ExPlanTech: Multi-agent system for intra-enterprise production planning and simulation based on ProPlanT technology.
Extra-Enterprise (EE) Agent: An agent operating in the extra-enterprise level of the system.
Private Knowledge: Information owned and administered by an individual itself. The private knowledge is not shared and can include collaboration preferences, restrictions, but also planning and scheduling algorithms and criteria.
Semi-Private Knowledge: Information shared within the members of a cluster or an alliance (e.g., availability of resources).
Public Knowledge: Information available to everyone in the system (e.g., enterprise name, address, and scope of interest).
Agent: Autonomous software or hardware entity loosely coupled into a multi-agent system. Operation of an agent-based system is based on collaborative (or sometimes self-interested) interactions of agents.
ExtraPlanT: Multi-agent system for extra-enterprise cooperation and collaboration by proactively communicating external agents based on ProPlanT technology.
Agentification: Process of integration of software and hardware equipment already existing in the enterprise into a multi-agent system.
Enterprise-to-Enterprise (E2E) Agent: An agent that establishes extra-enterprise cooperation and collaboration by proactively connecting external E2E agents.
e-Cat: Concept and research prototype of a catalogue for advertising and searching competencies and profiles within a cluster of cooperating partners. It is based on hybrid peer-to-peer architecture supported by two centralized units ensuring identity of partners and common understanding of competencies that are advertised within the community.
Intra-Enterprise Level: The part of the system operating inside a company. It should include planning and scheduling internal activities, simulation, and product/process flows of the company.