Dispatching AGVs with noisy estimation of crane operation time

The operation of both quay cranes and stacking cranes used in a container terminal is dynamically affected by a variety of factors. For this reason, errors could arise regarding the estimation of the time involved in processing containers. These errors make it difficult to dispatch automated guided vehicles (AGVs) used for internal transportation of containers in the terminal. In this paper, we propose a simulation-based dispatching algorithm which reduces the impact of the noisy estimation of crane processing time. When dispatching an AGV, the proposed algorithm collects stochastic samples of the evaluation value for each dispatching option by conducting multiple stochastic simulations of the AGV operation. Based on such samples, the proposed algorithm compares the dispatching options and selects the best one. For the stochastic simulation, we use a simple noise model of crane processing time. The effectiveness of the proposed algorithm is validated by simulation experiments.