Solving cost based robotic assembly line problems using variants of particle swarm optimization

In today´s competitive world, reducing the cost of the manufacturing component of production is on the mind of manufacturers all over the globe. In a manufacturing scenario, assembly is one of the major component. Robots are extensively used in an assembly line instead of manual labor. Robots eliminate the costs associated with manual workers - in terms of wages, training, health and safety, holidays and employee administration. Robots help to reduce both direct costs and overhead costs. In this work, a new robotic assembly line balancing (RALB) problem is formulated with an objective of minimizing the total production cost of an assembly line by allocating tasks to the workstations and assigning the cost efficient robot available. A recursive allocation procedure is used to assign the tasks to the workstations and assign the best available robot based on cost. Due to NP-hard nature of the problem four variants of an evolutionary algorithm is proposed to find the optimal solution of RALB problem. The performance of the proposed algorithm is tested and compared for small and large size benchmark problem instances and it is reported that PSO variant with varying constriction factor performs better in most of the datasets in terms of the quality of the solution.