Line balancing for mixed-model with deterministic and stochastic task times

Assembly line balancing for single model is to assign work tasks to stations such that the idle time. For mixed-model assembly line, the balancing objective is not only to minimize the idle time, but also to reduce the variation of assignments.;In this research, the objective is to develop heuristics for mixed-model line balancing problem with deterministic and stochastic task times. It is assumed that the same tasks in different models be assigned to the same station. The variation of assignments is evaluated using weighted sum of squares of the deviation of assignments from model cycle time or common cycle time. Maximal feasible solutions are found for each station to reduce the number of feasible solutions. It will also reduce the number of stations.;Two heuristics are developed to balance the mixed-model assembly line. Method (1) chooses one solution with the minimum variation from all or pre-determined number of maximal feasible solutions at each station.;Method (2) chooses two semi-final solutions with the least variations for each station following their path. The two final solutions are chosen from the four semi-final solutions based on the total variation. The final solution for the mixed-model line is either the one with the least total weighted sum of squares if both solutions include all tasks or the solution which includes all tasks.;For stochastic task times, both heuristics will adjust the station time using the mean and standard deviation of station time.;The heuristics are evaluated by randomly generated problems and a case study. On average method (2) results in fewer stations than method (1). Method (2) also results in less variation of assignments in all levels of factors.;Both procedures have been programmed using TURBOBASIC. Large problems can be solved. Different solutions can be obtained by changing the number of maximal feasible solutions at each station. For stochastic task times, it will result in a solution with certain probability that the assignments will be finished within the cycle time on average.