AN AGENT-BASED SIMULATION MODEL TO INVESTIGATE THE PRICING OF RURAL ROADS IN SASKATCHEWAN

Abstract

Industry in Saskatchewan, including natural resource, manufacturing and agriculture, is dependent on road infrastructure to reach suppliers and markets. The 296 Rural Municipalities (RMs) in Saskatchewan are responsible for the construction and provision of the extensive rural road network consisting of mostly gravel roads, which can be costly to maintain under heavy vehicle traffic. In general, road users do not directly pay the road provider for their road use; however, the decisions of road users can affect the costs incurred by road providers and vice versa. The goal of this research was to determine the feasibility of applying an agent-based model (ABM) to represent and compare the road use and road provision of a rural road network in Saskatchewan. The main objective was to develop an ABM to determine whether pricing a rural road network on an incremental cost basis would result in a net benefit when considering combined road use and road provision costs. The developed ABM included: road segments, nodes (intersections), road users, and a road provider. Simple heuristics were used to represent road use and road provision decision making, including least cost routing and traffic-based prioritization for road upgrade decisions. Vehicle traffic in the model was generated based on exogenous origin-destination (OD) inputs. The ABM was developed using a hypothetical road network and was applied to a case study rural road network. The permit fees considered in this research essentially shifted the incremental costs for road provision under traffic loading from the road provider to the road users. The purpose of this type of permit fee structure was to investigate road management policies that may be more cost effective when considering combined costs (road use and road provision costs). This type of permitting could provide a more direct linkage between road use and payment to fund road provision, which may be more equitable than current road funding mechanisms (e.g., gas tax, property tax). Model runs showed that the inclusion of permit fees incentivized road users for some OD pairs to change their routing and drive longer distances in order to drive larger percentages of their routes on upgraded road segments. This change in routing caused road user costs to increase with longer distances driven, and road provision costs to decrease due to lower traffic on gravel road segments. The shift in traffic routing due to the inclusion of permit fees was also found to change the road segments selected for upgrade, based on the simple traffic count prioritization upgrade criteria. While each considered scenario resulted in a net benefit (reduction in road provision and road use costs), the magnitude of the net benefit was consistently marginal in relation to the base case (without permit fees). Since permit fees were based on the incremental road provision cost, and, generally, road provision costs were small relative to road use costs, the magnitude of the permit fee does not impose strong incentives for altering road user behavior. Permit fees only altered road user route choice if there were alternative routes in which the road users did not have to significantly increase their trip distance (and costs) to find routes with upgraded road segments. Given the relatively low impact of permit fees on resulting combined costs found in the scenarios considered, the associated administrative costs may not be worthwhile for a road provider to implement such a permit fee structure. A sensitivity analysis was completed for select parameters of the model. The analysis provided insights into the selection of road provision parameters resulting in the lowest road provision and road use costs such as: minimum traffic levels used for road upgrade criteria, annual budget levels, and the impact of shifting traffic patterns. The model developed in this research illustrates the feasibility of using an ABM to support decision making involving road use and road provision policies. The complexities involved in road use, road provision and road performance required several simplifying assumptions to complete the model. Nonetheless, the results produced with the model illustrate the potential implications of various road use and road provision decisions. Further work to expand simplifying assumptions and refine model inputs may allow the model to become useful for road providers in understanding the impacts of alternative road provision policies for real world road networks.