Efficient approaches for robotic assembly line balancing problems

Assembly Line balancing (ALB) problems deal with the allocation of the tasks among workstations such a way that the precedence relations are not violated and a given objective function is optimized. It is a fundamental problem in continuous production line, and it is one of the difficult optimization problems. Installing assembly line is a long-term decision and required high capital investments. Hence, it is very important to design the assembly line and balance the workload on the workstations. The assembly line has to be rebalanced periodically or if there is a change in the production plan or process. Based on the strategic goals of the manufacturers, the performance measures have to be carefully chosen, since balancing decisions have a long term effect. Due to the technological advancements, human workforce is replaced by robots to perform the tasks in an assembly line. Different robots with different capacity and specialization are available to perform the assembly task, hence it is required to choose the best fit robot among the available robots such a way that it helps in improving the productivity of the assembly line. Robotic assembly line balancing (RALB) problem aims at assigning the tasks to workstation and allocate robot for each workstation in such a way that the productivity is improved. Very few researchers have proposed models for balancing a robotic assembly line. The main objective of this research is to develop efficient algorithms to solve robotic assembly line balancing problems. RALB problem is NP-hard, since the basic version of assembly line balancing problems falls under this category. To solve problem of this nature it is necessary to use metaheuristic algorithms. RALB problems with different objective functions are proposed and solved. The objectives considered for the RALB study are: minimizing cycle time, minimizing energy consumption, minimizing assembly line cost and maximizing line efficiency of a robotic assembly line. Straight and U-shaped RALB problems are considered. The results obtained for the two assembly line problems are compared. RALB problem with an objective of minimizing cycle time is solved using Particle Swarm Optimization (PSO) and hybrid models of PSO and efficient metaheuristics. Two allocation procedures are used for allocating tasks and robots in the assembly line. PSO and its variants are the metaheuristics proposed to solve the RALB problem. PSO is also hybridized with Genetic Algorithm and Cuckoo search to solve RALB problem. Proposed algorithms are able to produce better results when compared with the benchmark results reported in the literature. Manufacturing industries give importance to the reduction of energy consumption due to the increase in energy cost and to create an eco-friendly environment. Due to the importance of reducing energy consumption in an assembly line, an energy based RALB problem is proposed. RALB problem with an objective of minimizing energy consumption for straight and U-shaped robotic assembly line is proposed. Particle swarm optimization algorithm is the metaheuristic used to solve the proposed model. Cost reduction is one of the important tasks for the manufacturing companies throughout the world. In this thesis, a cost based RALB problem is also proposed. The objective considered in this problem is to minimize the assembly line cost. Particle swarm optimization and Differential evolution algorithms are proposed to solve the problem. Straight and U-shaped robotic assembly line problems are solved using the proposed algorithms and the results obtained are presented Since the investment in assembly line is high, industries try to maximize their usage in the shortest time possible. Maximizing the line efficiency is another measure considered in this thesis. Particle swarm optimization and Differential evolution algorithms are proposed to solve this RALB problem. Line efficiency of both straight and U-shaped configuration of robotic assembly line is compared. The research on RALB problem optimizing various performance measures considered reveals that U-shaped robotic assembly line is better than straight robotic assembly line.