Determining the optimum toll level for freight vehicles in urban conditions

© 2019 Hewage Loshaka Kumara PereraRoad pricing has been a very common practice in the world for many decades. Tolls are charged from road users as a tool to control traffic (congestion) or to raise finance (to recover capital and maintenance cost) or to control emissions. Public-Private Partnerships (PPP) or Build-Operate-Transfer (BOT) schemes have become common sources of financing infrastructure projects in many countries and tolls are collected to recover the cost. Private investors are attracted to such investments due to the potential opportunity to make high returns. On the contrary, the public has accepted private investments due to its transparency and rapid nature of construction compared with government authorities. However, in the operational stage, the public is more concerned about factors such as air quality, safety, and noise whereas investors are mostly concerned about the return on investment, which are in conflict. Nevertheless, toll levels appear to be set for maximizing profits ignoring many public concerns and other related issues. Road prices are charged usually to recover pavement damage costs, road maintenance costs, capital costs, congestion costs, marginal savings, and emission costs. Regardless of the intended objective of charging vehicles on road, many studies in the past have looked at how various road pricing policies would change the behaviour of car traffic. But limited studies have been carried out identifying its impacts towards freight user behaviour. Demand for freight movement is increasing day to day as populations’ increase, especially in urban cities, which makes them significant. There were only a few studies carried out in the past to study the behaviour of freight vehicles, especially how they respond to toll charges. It was found that toll avoiding nature of freight vehicles increases the social and environmental costs. City logistics is a concept which looks at total optimization of freight movement in city limits. City logistics consider system optimisation not limited to economic aspects, but also social and environmental. Freight vehicles are known to produce more externalities compared to cars, especially in city limits, where average speeds are low and most congested. As a result, freight vehicles in urban conditions produce more externalities such as crashes, noise, emissions, and it was found that conditions are getting worst due to toll avoiding nature of freight movement. Determining optimum toll charges for freight vehicles are crucial decisions for policymakers considering socio-economic aspects. Therefore, the objective of this study is to develop a mathematical approach to design an optimal toll scheme for multi-class vehicles, considering direct costs and externalities. A bi-level modeling approach is used, where toll prices for multi-class vehicles are decided in the upper level and user response to toll charges are predicted using user equilibrium (UE) conditions with multi-class traffic assignment in the lower level. In the upper level of the model development total cost calculation is an important component and thus a freight cost model was developed for Australia. For the cost model, local data were used when possible and gaps were filled from literature. In the lower level, the response to toll charges was embedded to user equilibrium conditions. A discrete choice experiment was carried out to find out how freight vehicles would respond to toll charges in Melbourne. This NP-hard, non-convex problem with non-linear constraints was solved using multi-objective optimisation. A Non-dominated sorting genetic algorithm was used for optimisation. The model is applied to a small hypothetical road network and a real network in Melbourne with static demand conditions. Several applications of the model were performed considering various stakeholder objectives and optimal solutions were identified for enhancing City Logistics. As an application, an effective toll charging scheme to minimize total network vehicular emissions while maintaining reasonable revenue for investors was found using the bi-level modeling approach. The model takes into consideration both toll revenue and total vehicle operating costs to produce an acceptable solution for both investors and road users. All major components of hazardous emissions (CO2, PM2.5, SO2 and NOx) were considered and estimated using an emission costs model developed based on secondary data quantifying their impact on human health and the environment. Application of the model was illustrated using two networks mentioned above. Since this is a single objective optimisation, genetic algorithm was used to find near-optimal solutions for each toll-charging scenario. The results revealed that commonly used toll schemes are inefficient with respect to multi-stakeholder objectives. The toll charging scheme optimized using the GA was able to reduce the total emission cost of the network by 12% compared to a toll scheme currently used by a toll facility in a real network in Melbourne, Australia. In the next section of this study, the toll problem is viewed from a different angle. As mentioned before toll roads constructed under public-private-partnerships (PPP) is very common around the world and due to the high capital cost invested and high risks associated with subsequent returns, investors are extremely concerned about future returns. As a result, the present trend is to charge high tolls for freight vehicles, especially in urban areas, which has created many problems as mentioned above. Considering the overall impacts, this section of the study identifies a tolling scheme with subsidies as the best option to minimize the total cost of urban freight transport which can lead to more sustainable city logistics in the future. A step by step process is followed in this study with illustrations to explore the best toll charging scheme considering various toll scenarios within a multi-objective framework. Acceptability of toll schemes found in this study by multi-stakeholders is a critical decisive factor determining successful implementation of toll schemes. Multi-Agent Multi-Criteria Analysis was used to evaluate the various toll schemes produced from this study under various conditions. Surveys were carried out to collect the required information to feed the MAMCA. Finally, the best toll alternatives are ranked based on MAMCA outcome which can be used by policy decision makers considering the given conditions