Hybrid multiobjective genetic algorithm for integrated dynamic scheduling and routing of jobs and automated guided vehicles in flexible manufacturing systems

• Main Author: Umar, Umar Ali
• Format: Thesis
• Language: English
• Published: 2014
• Online Access: http://psasir.upm.edu.my/id/eprint/64738/1/FK%202014%20162IR-unlocked-converted.pdf; http://psasir.upm.edu.my/id/eprint/64738/

The dynamic continues trend of adoption and improvement inventive automated technologies is one of the main competing strategies of many manufacturing industries. Effective integrated operations management of Automated Guided Vehicle (AGV) system in Flexible Manufacturing System (FMS) environment results in the overall system performance. Routing AGVs was proved to be NP-Complete and scheduling of jobs was also proved to be NP hard problems. The running time of any deterministic algorithms solving these types of problems increases very rapidly with the size of the problem, which can be many years with any computational resources available presently. Solving AGVs conflict free routing, dispatching and simultaneous scheduling of the jobs and AGVs in FMS in an integrated manner is identified as the only means of safeguarding the feasibility of the solution to each sub-problem. Genetic algorithm has recorded of huge success in solving NP-Complete optimization problems with similar nature to this problem. The objectives of this research are to develop an algorithm for integrated scheduling and conflict-free routing of jobs and AGVs in FMS environment using a hybrid genetic algorithm, ensure the algorithm validity and improvement on the performance of the developed algorithm. The algorithm generates an integrated scheduling and detail paths route while optimizing makespan, AGV travel time, mean flow time and penalty cost due to jobs tardiness and delay as a result of conflict avoidance. The integrated algorithms use two genetic representations for the individual solution entire sub-chromosomes. The first three sub-chromosomes use random keys to represent jobs sequencing, operations allocation on machines and AGV dispatching, while the remaining sub-chromosomes are representing particular routing paths to be used by each dispatched AGV. The multiobjective fitness function use adaptive weight approach to assign weights to each objective for every generation based on objective improvement performance. Fuzzy expert system is used to control genetic operators using the overall population performance history. The algorithm used weight mapping crossover (WMX) and Insertion Mutation (IM) as genetic operators for sub-chromosomes represented with priority-based representation. Parameterized uniform crossover (PUX) and migration are used as genetic operators for sub-chromosomes represented using random-key based encoding. Computational experiments were conducted on the developed algorithm coded in Matlab to test the effectiveness of the algorithm. First scenario uses static consideration, the second scenario uses dynamic consideration with machine failure recovery. Sensitivity analysis and convergence analysis was also conducted. The results show the effectiveness of the proposed algorithm in generating the integrated scheduling, AGVs dispatching and conflict-free routing. The comparison of the result of the developed integrated algorithm using two benchmark FMS scheduling algorithms datasets is conducted. The comparison shows the improvement of 1.1% and 16% in makespan of the first and the second benchmark production dataset respectively. The major novelty of the algorithm is an integrated approach to the individual sub-problems which ensures the legality, and feasibility of all solutions generated for various sub-problems which in the literature are considered separately.