A negotiation scheme for autonomous agents in job shop scheduling

Owing to the increasing impact of some environmental factors on manufacturing firms' performance, in the past few years growing attention has been paid to the reorganization and optimization of operations strategies. The complexity of manufacturing processes has increased due to factors such as demand variability and unpredictability, shorter product life cycles, market globalization and competitiveness increase. A modern manufacturing system is required to be both sufficiently reactive and reconfigurable and, at the same time, well performing in stable and predictable conditions. The achievement of these objectives often requires the implementation of adequate production control systems. While traditional scheduling and planning approaches propose a centralized system producing a unique production plan for all machines and production lots, multi-agent control architectures are based on the idea that several distributed decision makers can cooperate and interact to obtain globally optimal performances. In this paper, we exploit an autonomous agent control system architecture that is slightly different from the ones in the literature, discuss its implementation details and report computational experience.