Optimization tool for transit bus fleet management

ransit agencies face the challenge of being environmentally-friendly, while maintaining cost-effective operation. Many studies have focused on investigating new bus technologies to reduce emissions and cost. However, they ignored the potential environmental and economic gain by improving the fitness and harmony between individual buses and routes in fleets. This dissertation provided, for the first time, a tool for fleet operator to intelligently dispatch buses and select new technologies that are tailored to their needs and business.;One key element in this tool is a bus life cycle cost model that can simulate and predict every capital and operational cost category for different bus technologies. The cost model was funded by Transportation Research Board and developed in Transit Cooperative Research Program (TCRP) C-15 project, the purpose of which was to assess hybrid-electric bus performance in real-world operation. The research team (author as a key member) picked four bus transit agencies among a handful of test sites that were operating hybrid-electric buses and collected 28 month bus operation data at almost all data collection sites. The sophisticated life cycle cost model is the backbone of this tool to calculate cost.;The other key element is a green house gas (GHG) emissions model, which was based on the fuel consumption model in the TCRP C-15 project and GREET model generated at Argonn National Laboratory. The GHG model utilizes fuel consumption data to provide tail pipe GHG emissions and well-to-tank GHG emissions for specific fuels and bus propulsion technologies.;The last key element is the use of genetic algorithm (GA) as a search and optimization scheme in the fleet management tool. A ranking matrix was developed to rate and compare different dispatch strategies on multiple criteria, which can vary in units or scales. When a fleet has large number of buses (dozens to thousands) on multiple routes, the number of all possible bus dispatch strategies becomes tremendously huge and difficult to explore. The GA uses the evolution theory of "Only the strongest survive" to find the best strategy.;The tool shows that optimization objectives dictated dispatch strategies that are successful in specific applications. For example, in a 35-bus fleet examined in this dissertation, the proposed dispatch strategy could reduce fleet well-to-wheels (WTW) GHG emissions up to 364 metric tons of carbon dioxide equivalents, a 17.5% GHG emissions reduction from the initial dispatch strategy. The same dispatch strategy increased...