Abstract
The mixed-model assembly line is a popular manufacturing system that produces various models in an intermixed sequence to minimise work overload at stations. However, the system dynamic of processing time variations might cause increased work overload during actual production of the predetermined model sequence. To reveal the critical stations in resulting in work overload increases, this research develops an improved global sensitivity analysis method that guides efficient production simulations of mixed-model assembly lines with randomised processing times and evaluates the impact degrees of different stations on the work overload value using Monte Carlo algorithm. In the case study of a diesel engine assembly line, the efficiency of the proposed method over traditional ones is demonstrated, and some insights about critical stations are generalised based on the impact degrees of various production plans.