Modeling and simulation of human behavior impact on production throughput

The development of new technologies generates intelligent, complex, and collaborative
production systems. Several research works want to improve the production performances
while improving the comfort of the human operator. However, it is not obvious to define optimal
strategies of operations planning and control that consider the unexpected and variable character
of human operators. It is necessary to understand and model human behavior to develop
predictive and dynamic actions. Even if some generic behaviors have been well integrated into
classical quasi-deterministic models, there is still a need to develop stochastic models even closer
to human behavior to allow more dynamic and anticipatory decision-making, especially at the
operational level. In this work, we propose to model human behavior by a Markov chain and
to evaluate the effect of the different behavior types on the production system performance. A
heterogeneous set of human operators, with different behavioral patterns, were generated and
tested through simulation. Earlier results demonstrate that there is a direct link between the
behavior of human operators and the performance of the production system. It demonstrates also
how to integrate such models in a dynamic decision-making process concerning, the assignment
of workers to workstations.