A multi-agent system for the integration of process planning and scheduling using operation-based time-extended negotiation protocols

  • Authors:
  • Izabel Cristina Zattar;Joao Carlos E. Ferreira;Joao Gabriel G. G. Rodrigues;Carlos Humberto B. de Sousa

  • Affiliations:
  • Universidade Federal do Parana, Setor de Tecnologia, Curitiba, PR, Brazil;Universidade Federal de Santa Catarina, Depto de Engenharia Mecanica, Florianopolis, SC, Brazil;Universidade Federal de Santa Catarina, Depto de Engenharia Mecanica, Florianopolis, SC, Brazil;Universidade Federal de Santa Catarina, Depto de Engenharia Mecanica, Florianopolis, SC, Brazil

  • Venue:
  • International Journal of Computer Integrated Manufacturing
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper proposes the on-line adaptation of a process plan with alternatives, through the application of an operation-based time-extended negotiation protocol for decision-making for the real-time routing of job orders of parts composed of machining operations in a job-shop environment. The protocol is modified from the contract net protocol, which specifies the interaction between agents in a multi-agent system, in order to cater for the multiple tasks and many-to-many negotiations. A multi-agent system model was developed in order to make decisions based on a heterarchical structure, which allows grouping manufacturing parts composed of machining operations through the use of the new protocol. The grouping of the machining operations enables reduction of set-up times, resulting from the reduction of machines changes in flexible manufacturing environments. For each part, all feasible routings are considered as alternative process plans, provided the different manufacturing times in each machine are taken into account. This study evaluates four types of flexibility-operation flexibility, sequencing flexibility, processing flexibility, and machine flexibility-to identify which can provide better performance. The time-extended negotiation period allows the visualisation of all of the times involved in the manufacture of each part, including those times that are not considered in systems of this nature, such as the negotiation times among agents. Extensive experiments have been conducted in the system, and the performance measures, including makespan and flow time, are compared with those obtained by the search technique based on the co-evolutionary algorithm.