(The) task allocation problem for automated assembly line systems

Abstract

The automated assembly has been one of the most important production means of the modern industry. Especially in the electronics and part production industries where products are small enough to be handled by a moderate-sized assembly robot, the automated assembly lines are gaining popularity. The line setup problem frequently arises in the design phase of these systems. The main decision variables of the line setup problem include the number of work stations and the corresponding task allocation plan. The objective of this dissertation is to develop an efficient solution approach to solve the task allocation problem for the automated assembly line systems. When the system is reliable the problem is reduced to a line balancing problem. For the purpose of line balancing in the automated assembly line system, we need to consider explicitly the parts and tools as well as the tasks as the objects for assigning. When a task is assigned to a robot cell, the part and tool needed to perform the task should also be assigned to that station. In addition to the constraints found in the simple assembly line balancing problem, the tray capacity constraints and slot capacity constraints must be satisfied by every feasible assignment. We develop two independent solution approaches to solve this problem. One is the cutting plane approach, and the other is the branch and bound algorithm which uses the dynamic upper bound. When the assembly robots are unreliable the completion time of each processing cycle is no more deterministic. The pure processing time is deterministic, but when the robot malfunctions in a normal processing cycle and stops operation, the required clear time until it resumes the remained operation is random. In this case the system should be treated as a stochastic one. For stochastic serial production systems, many researchers pointed out that balancing workloads over stations is not optimal. They showed a better production rate could be achieved by delibe...