Simulation-Based Optimization Method for Impact Evaluation to Work Zones in Large-Scale Networks

Abstract

Work zones for road maintenance in traffic networks can significantly impact the traffic distribution and route choice behaviour of travellers. This study proposes an approach to evaluate and predict the broad-scale effects of work zones on large-scale traffic networks. For the requirement of the efficient evaluation of the various impacts of work zones on traffic networks, this study defines the road maintenance sensitivity factor (RMSF) to represent the joint impact of work zones. A simulation-based optimization method for calibrating the RMSF is formulated. The original objective function is replaced by an analytical metamodel that builds the approximate relationship between the RMSFs and traffic flow distribution with the effect of work zones. A derivative-free trust-region algorithm is used to obtain the optimal solution. Numerical experiments are conducted on a small test network and a large-scale freeway network. The proposed method shows the accuracy and effectiveness with tight computational resources than the simultaneous perturbation stochastic approximation method in both experiments, giving the RMSF results and map the traffic redistribution of large-scale networks with work zones accurately and efficiently, which can help traffic managers to optimize maintenance plans and traffic management measures with the assistance of the traffic management system.