Abstract:
This thesis presents the parallel machines scheduling problem withsequencedependent setup time in the tire buildingprocess by using the metaheuristic. The objective is to design a metaheuristic method of production scheduling and calculate the total cost of the production to be appropriately low that can be used in the real world. In addition, to provide more options for the methods used in productionscheduling. In which this thesis proposes a Genetic Algorithm (GA) for parallel machine scheduling. The total production cost is set to be the objective function of the finding production schedule process by genetic algorithm. The proposed GA used a set of random keys as a chromosome that represents the sequence of jobs performed on their assigned machine. Three genetic operations were used for transforming the population to the next generation including elite reproduction, parametric uniform crossover, and randomized mutation. Defining a maximum result processing time of 550 seconds. The proposed method was tested by using 30 dailyscheduling problems from an industry with 71-86 job models and 32 parallel machines. The result of efficiency compared with mixed-integer linear programming (MILP) method. Which the proposed methods increased the total cost of production average by 2,935.8 baht per day(or 4.1%). But in terms of the time of the maximum result processing that reduced950 seconds (or 63.3%), including there is nocost to use the processing tools that reduced420,000 baths (or 100%). So this thesis can make the parallel machines scheduling by using the metaheuristic method tocalculate the cost appropriately low that can be used in the real world andadd an alternative to production scheduling by purposes efficiently.