Applicability of agent-based model to managing roadway infrastructure
MetadataShow full item record
In a roadway network, infrastructure conditions determine efficient network operation and traveler safety, and thus roadway engineers need a sophisticated plan to monitor and maintain network performance. Developing a comprehensive maintenance and rehabilitation (M&R) strategy for an infrastructure system, specifically a roadway network, is a complicated process because of the system uncertainties and multiple parties involved. Traditional approaches are mostly top-down, and restrict the decision-making process. In contrast, agent-based models, a bottom-up approach, could well simulate and analyze the autonomy of each party and their interactions in the infrastructure network. In this thesis, an agent-based model prototype was developed to simulate the operations of a small roadway network with a high degree of simplification. The objective of this study is to assess the applicability of agent-based modeling for infrastructure management problems through the following four aspects: (1) to simulate the user route selection process in the network; (2) to analyze the impact of users’ choices on the congestion levels and structural conditions of roadway sections; (3) to help the engineer to determine M&R strategies under a certain budget; and (4) to investigate the impact due to different fare rates of the toll road section on the infrastructure conditions in the network. This prototype detected traffic flow, and gave appropriate M&R advice to each roadway segment. To improve this model, more investigation should be conducted to increase the level of sophistication for the interaction rules between agents, the route selection, and the budget allocation algorithm. Upon completion, this model can be applied to existing road networks to assist roadway engineers in managing the network with an efficient M&R plan and toll rate.