Uma abordagem multi-agente para competição de agentes negociadores em gerenciamento de cadeias de suprimento. / A multi-agent approach to negotiating agents of competition in supply chain management.

AUTOR(ES)
FONTE

IBICT - Instituto Brasileiro de Informação em Ciência e Tecnologia

DATA DE PUBLICAÇÃO

22/04/2010

RESUMO

This work aimed to design a computational approach to supply chain management problem in the scenario Trading Agent Competition for Supply Chain (TAC-SCM) when disturbances on manufacturing process are considered. Simulations have shown that the decentralized approach based on smart products is promising in terms of robustness of the system of control and planning. However, this same approach has achieved poor results in terms of profit. Thus, it is required more research into how systems can make the best of both approaches: centralized planning and decentralized planning. To achieve a hybrid multi-agent system and robust to these problems, we investigated how the social regulation of the market, intelligent products and machine learning could be combined as a mechanism of distributed control of resources and responsibilities. The result showed that the solution developed is promising when combined with techniques appropriate machine learning (or even good heuristics) and some market mechanism controlling the distribution of resources for distributed elements of the system.

ASSUNTO(S)

sistemas multi-agente inteligência artificial produtos inteligentes ciencia da computacao multi-agent system artificial intelligence intelligent products

ACESSO AO ARTIGO

Documentos Relacionados