Multi Agents approach for Job Shop Scheduling Problem using Genetic Algorithm and Variable Neighborhood Search method

The job shop scheduling is an important and complex problem for a manufacturing system. It is well known and popular problem having (Non-Polynomial) NP-hard characteristic to find the optimal or near optimal solution (schedules) quickly. In job shop scheduling a set of “N” number of jobs are processed through an “M” number of given set of machines. It must be processed in the prescribed order by utilizing the feasible sequence of operations for a job. Therefore, due to its complex nature the finding of approximate solutions are chosen rather than the finding exact solution which involve higher cost. Various meta-heuristics techniques are utilized in order to find the sub-optimal solution for job shop scheduling problem. Genetic Algorithm (GA) and Variable Neighborhood Search (VNS) method are the preferred techniques which are bestknown for global and local search of solutions respectively. VNS is work as in augment for GA approach. In the present work, Multi-agents are proposed to find the near optimal solution for job shop scheduling problem using GA and VNS approach in parallel. Multi-agent system is preferred owing to its ability to perform in parallel and robustness as well as the elucidation of intelligence. In the proposed system, many hosts of the network are accommodated with agents. JADE is used to setup communications. Each agents is designed to perform the specific task namely Initialization Agent (IA), Processing Agent (PA) and Coordinating Agent (CA) for initial generation of population, to schedule the operations on machines, to find the distinctive host and to perform the migrations between various populations respectively. The objective is to find an optimal value of makespan for the job shop scheduling problem. The performance of the system is assessed by a case study and it reveals that the proposed approach is effective enough to find the optimal solution. Future works is to introduce Disturbance Agents (DA) for internal and external disruptions.