Abarja

Multi Agent Resource Scheduling System - VERSION 1.0 To see VERSION 0.0 go to github.com/granales/MARSS This program provides real time resource scheduling with short reaction times to disruptions such as machinery faults. Reaction times to reschedule becomes feasible taking minutes. This project does not aim to reduce uncertainty, it provides a tool to manage it. The dynamic scheduling auction-based multiagent system is thought to provide stability to a production facility as it can dynamically and periodically create schedules. This program conforms a decentralized decision-making system. It is developed to monitor production processes and rescheduling the orders if necessary. The main characteristic of the system is that it is based on the existing historical data. The system is thought to help the planners monitor the order production process and develop a rescheduling decision when the production is interrupted. The system generates a feasible solution in the short time available. In order to increase the feasibility and flexibility of such systems, agents are assigned with different roles. With the use of multi agent systems, the steel production scheduling problem becomes populated by several autonomous and heterogeneous intelligent agents with various goals, constraints, and capabilities. These agents can cooperate and coordinate their local schedules to generate desirable global schedules. Agents negotiate in auctions to assign materials to-be processed to processing resources.