Mixed-Model Assembly Line Balancing with Fractional Task Allocation under Multiple Demand Scenarios

Abstract

This study proposes a two-stage approach that sequentially solves two mixed-integer programming models for a mixed-model assembly line that accepts fractional task allocation under multiple demand scenarios. The task sequence is assumed to be fixed in all demand scenarios, but the number of stations and the allocation of tasks to stations can be changed according to the realized demand. Only one task from among all of the tasks can be shared between adjacent stations, and the shared task can be moved to the boundary of these stations. The upper mathematical model decides the task sequence and the allocation of tasks to stations in all scenarios. The selection of the shared task is also included in the upper model. The lower model generates a detailed assembly schedule for each scenario and allocates buffers between stations when necessary. Numerical experiments provided several findings as described below. On average, it was beneficial to design an assembly line so as to be prepared for all scenarios under multiple demand scenarios. Introducing task sharing expands the solution space and may reduce the number of stations used. However, responding to unexpected demand scenarios through task sharing is limited. The number of buffers in the line affects the desirable sequencing of product models.