Application of Genetic Algorithm to Mass Production Line for Productivity Improvement Dr. Poornima G. Naik Assistant Professor Department of Computer Studies Chh Shahu Institute of Business Education and Research, Kolhapur, India Girish R. Naik Associate Professor Production Department KIT’s College of Engineering Kolhapur,India Abstract— Genetic Algorithm (GA) is an invaluable tool for solving optimization problems due to its robustness. It does not break even if the inputs are changed slightly or in the presence of a reasonable noise. GA offers a significant benefits over other optimization techniques in searching a large state space or n-dimensional surface In this paper we have made an attempt to study the effect of population size cross over and mutation on the performance and convergence of GA. The criteria for adopting the proper selection method is also studied. There is lot of literature on application of GA in various domains but best to our knowledge there exist very few papers which discuss the distribution of population, criteria for choosing selection method. Finally, we use the results for optimizing the non-value added cost component of a cycle time of constrained resources for productivity improvement. The problem is for medium scale manufacturing plant and is solved using GA toolbox of MATLAB. The results are used in redesigning the assembly line to overcome the limitations offered by constrained resources. Keywords – Genetic Algorithm, MATLAB, Cycle Time, Constraint Resource, Production Line, Non-linear Optimization. Theory of Constraints I. INTRODUCTION Genetic algorithms (GA) are adaptive heuristic search algorithms based on the evolutionary ideas of natural selection and genetics. As such they represent an intelligent exploitation of a random search used to solve optimization problems [1]. Although randomized, GAs are inherently not random instead they exploit historical information to direct the search into the region of better performance within the search space. A space of all feasible solutions is referred to as search space. Each and every point in the search space represents one possible solution. Each possible solution can be marked by its fitness value and GA looks for a best solution among a number of possible solutions represented by one point in the search space. Keeping in memory more than a single solution at each iteration offers lot of advantages. The algorithm can span larger search space for better solutions [2]. It can recombine better solutions to get still better ones. In optimization a few random changes can be a good way of exploiting a search space quickly. The distribution of population over search space together with mutation prevents the algorithm to be trapped in a local minimum [3]. GAs are something worth trying when everything else fails or when we know absolutely nothing of the search space [4,5]. Production lines are flow line production systems which are of great importance in the industrial production of high quantity standardized products and more recently even gained importance in low volume production of customized products. Small and Medium Scale Enterprises (SME) have lack of knowledge to manage the company resources especially in the production line [6]. The challenge is to reduce the cycle time, thus reducing the lead time of production. Cost reduction and inventory reduction have become more important for survival of an enterprise. Theory of Constraints (TOC) is one such technique for productivity improvement which focuses on improving system performance [7]. TOC tries to identify constraints in the system and exploit and elevate them to improve the overall output of the system. However, there is a practical limitation on the value up to which a cycle time can be reduced. If the line balancing is not achieved within this practical limit alternative methods such as buffer stocks, shifting some of the operations to a new machine are to be considered [8]. International Journal of Latest Trends in Engineering and Technology (IJLTET) Special Issue - IDEAS-2013 125 ISSN: 2278-621X