Development and analysis of production distribution plans for a bearing manufacturing company to meet deterministic and probabilistic demands

Abstract

In today s supply chain network, many industries try to move closer to their customers by adopting distributed production approach. Hence, the generation of appropriate production-distribution plans gains importance. In this thesis, integrated aggregate production distribution plan considering three different types of production costs, viz., regular production, overtime production and outsourced production cost along with inventory holding cost, backorder cost, hiring cost, laying-off cost and trip-wise transportation cost is developed for a renowned bearing manufacturing industry in India. Also, in the later part of the thesis, the production-distribution plans are generated for multi-criteria deterministic and stochastic demand scenarios. An attempt is also made to solve the integer non-linear programming model using LINGO 8.0, one of the popular operations research software which works based on branch and bound algorithm. LINGO software results are used as benchmark solutions. Finally, integrated production distribution plan considering three major objectives, viz., total cost minimization, change in labour level reduction and under utilization minimization, is developed for the bearing manufacturing industry. Change in labour level reduction and percentage under utilization minimization are considered as other objectives apart from the total cost minimization objective. The multi criteria model is solved using a novel analytic hierarchy process heuristic discrete particle swarm optimization algorithm. The algorithm is used for obtaining the best production distribution plan for both stochastic and deterministic demand scenarios. The deterministic demand problems are solved by clubbing analytic hierarchy process heuristic discrete particle swarm optimization algorithm with simulation approach. In addition to the bearing manufacturing industry data set, two other test data sets are also solved. newline