Event-based optimization for dispatching policies in material handling systems of general assembly lines

A material handling (MH) system of a general assembly line dispatching parts from inventory to working buffers could be complicated and costly to operate. Generally it is extremely difficult to find the optimal dispatching policy due to the complicated system dynamics and the large problem size. In this paper, we formulate the dispatching problem as a Markov decision process (MDP), and use event-based optimization framework to overcome the difficulty caused by problem dimensionality and size. By exploiting the problem structures, we focus on responding to certain events instead of all state transitions, so that the number of aggregated potential function (i.e., value function) is scaled to the square of the system size despite of the exponential growth of the state space. This effectively reduces the computational requirements to a level that is acceptable in practice. We then develop a sample path based algorithm to estimate the potentials, and implement a gradient-based policy optimization procedure. Numerical results demonstrate that the policies obtained by the event-based optimization approach significantly outperform the current dispatching method in production.