Dear colleague,
There will be a talk in Operations Research given by Judith Wang (School of Civil Engineering and Institute for Transport Studies, University of Leeds) next 9 July 2024, from 14:00 to 15:30 Room: Seminar room, building VII.
----------------------------------------------
Seminar of Operations Research
----------------------------------------------
Title: Considerations of Sustainability in Transportation - A Case for Multi-objective Optimisation
Speaker: Judith Wang (School of Civil Engineering and Institute for Transport Studies, University of Leeds)
Date/time: 9/07/2024 from 14:00 to 15:30
Room: Seminar room - building VII
Abstract: Traditionally, optimisation models used in transportation planning consider the optimisation of economic objectives such as minimising travel distance, travel time and monetary cost or maximising the more abstract concept of "utility". However, transportation is not a purely economic activity. It affects people and the environment in many ways, e.g., through fuel consumption and pollution. Hence decision makers at all levels - from a worker choosing a route for his commute to work to a national government aiming to reduce greenhouse gas emissions - are increasingly interested to pursue aims that fall under the wide umbrella of sustainability when making transport decisions. In this talk, we present four examples that illustrate how multi-objective optimisation can be a valuable tool to support decision making for sustainable transport.
Keywords: Transportation, Health, Reliability, Sustainability, Shortest Path, Traffic Assignment, Bi-level Optimisation, Multi-objective Optimisation
Healthy Route Choice for Commuter Cyclists
Commuter cyclists are often motivated by health benefits of cycling, yet they are vulnerable to exposure to traffic-related air pollution. Hence, in their route choice they face trade-offs between travel time and pollutant dose. We develop a bi-objective shortest path model, which considers the minimisation of travel time and pollutant dose. Travel time takes into account the length and slope of each link on a path as well as the average speed of the cyclist. Pollutant dose is calculated based on traffic flow, speed and vehicle fleet composition. Using a vehicle emissions prediction model these are converted to pollutant concentrations and multiplied by travel time and cyclist minute ventilation rate finally result in pollutant dose values for each link. Since both travel time and pollutant dose are additive, we can employ a bi-objective label correcting algorithm to find the set of efficient paths from which the cyclist can choose. The model is applied in a case study in Auckland, New Zealand.
Walking School Bus Line Routing for Efficiency, Health and Walkability
Walking School Bus (WSB) has been recognised as an innovative solution to promote walking to school, bringing a wide spectrum of benefits, including: health benefits from the physical exercise, social skills and traffic reduction. To facilitate the success of WSBs, one vital element is its route planning, which directly affects the catchment for the service and the realisation of all the potential benefits. Previously, time has been the only factor that has been considered in WSB routing problems. Other important factors including air quality, safety and comfort will also be considered in this paper. Air quality along a WSB route is important to help realise the health benefits of walking. Traffic safety has been the biggest barrier to walking to school and must be addressed in planning a WSB route. Ensuring children have an enjoyable and comfortable experience is vital for the sustainability and success of WSB.
A walking network is introduced to enable modelling pedestrian movements in detail, including walking movements on different sides of the road and crossing movements. This approach enables detailed route-based analysis to assess the localised effect of air quality on pollutant dose. We define walkability as a measure of children’s needs in safety and comfort, which can also be assessed in detail on each route. We propose a multi-objective optimisation model to generate efficient WSB routes with three objectives representing the potential benefits of WSB: (1) to minimise time; (2) to minimise pollutant dose; and (3) to maximise walkability. Our multi-objective WSB route planning model is highly transferable to any selected school in any WSB targeted area. Planners will be able to select a combination of WSB lines to offer, based on the requirement of coverage area and resource availability.
Bi-objective User Equilibrium Models of Travel Time Reliability
Models of route choice are also central to user equilibrium models in traffic assignment. Traffic assignment is concerned with determining how traffic flow satisfying demand for a certain number of trips between origins and destinations in a road network is distributed on the links of this network. This results in equilibrium models, where at equilibrium no user (driver) has an incentive to unilaterally switch routes, because all traffic is on "minimum cost'' routes. So what does "minimum cost'' mean? Empirical studies have shown that the three most important factors that influence route choice behaviour of drivers are travel time, travel time reliability and monetary cost. In many models considering these factors, a generalised cost function making use of values of time and/or values of reliability (possibly for different user classes) is considered as the route choice function, i.e., it is assumed that all drivers aim to minimise their generalised cost when choosing a route from their origin to their destination. In contrast we introduce a definition of bi-objective user equilibrium.
We show that the earlier concepts of travel-time-budget user equilibrium and late-arrival-penalty user equilibrium are special cases of this definition. In fact, the definition of bi-objective user equilibrium gives rise to a set of TTR-BUE flows, and generalised cost user equilibrium flows are specific TTR-BUE flows.
Road Pricing for Sustainability
Congestion pricing is a policy instrument that is used in many cities around the world, e.g., Singapore, London and Stockholm, to reduce congestion of the road network during peak hours. But it is also part of wider strategies to enhance sustainability in transport. It can improve the environment in terms of air quality and hence reduce the negative impact of vehicle emissions on health. From an optimisation point of view, to maximise the effectiveness of congestion pricing, it is only natural to consider internalising the external costs of air pollution, including costs associated with their impact on the environment and population health, by charging road users an appropriate toll.
We consider a bi-level multi-objective model to achieve this goal. The upper level models the decision making process of the policy decision makers, which are to minimise total travel time, to minimise total vehicle emissions and to minimise negative impact on health modelled as, for example, median population CO dose. Thus at the upper level we consider a three-objective optimisation problem to determine link tolls for all links in the network.
The short bio of the author can be found attached.
Best regards,
Graça Gonçalves and Lídia Lourenço
Department of Mathematics
Center for Mathematics and Applications
NOVA SCHOOL OF SCIENCE AND TECHNOLOGY | FCT NOVA
Campus de Caparica | 2829-516 Caparica | Portugal
