Study of relocation strategies for electric car sharing system

In order to see how different relocation strategies including relocation for charging and relocation for balancing the demand. I use a simulator to simulate real traffic situation in the city of Turin in Italy. I generate thousands of real FFCS trips in given time period. Firstly, by posing three different kinds of charging relocation strategies that choose charging poles when finishing trip, I observe their performance and study the which additional cost which relocation operation bring. Post charging relocation are also called reactive relocation in the thesis. Next, my partner and I consider ‘proactive’ relocation. In that way, relocation is not only happened at the end of each trip, but can be scheduled and operated in a given time frequency. We use Kernel Density Estimation (KDE) to catch the demand spatial variability. We propose an hourly triggered relocation strategy. We relocate the cars to some zones which are confronted with the KDE model at the end of each hour. By changing the fleet size and the number of relocation workers, we analyze many Key Performance Indexes such as satisfaction fraction of booking trip, relocation cost and system revenue so as to consider the availability of this algorithm. The main questions that we try to answer are: •??Can I add more configuration to the existing simulator, able to implement different kinds of both reactive and proactive relocation strategies? •??Can I fairly compare system performance of different relocation strategies? •??Can I fairly compare financial performance of different relocation strategies? More specifically, the research questions I pose to analyze are the following: •??How do closest-free/closest-queueing/random post charging relocation strategies influences service quality and operational cost in the given city Turin? •??How do proactive relocation strategy influences service quality and operational cost comparing with no relocation strategy scenario in the given city Turin? •??How do system parameters such as fleet size ,number of relocation workers and charging poles density impact service quality and operational cost in the given city Turin? Our results show that in the reactive model, taking the low battery car to the nearest available charging poles has the most efficient performance. Besides, proactive relocation strategy make the whole system maximize the satisfied demand by increasing relatively acceptable additional relocation cost. The thesis is organized as follows: In Chapter II I propose more detailed introduction about previous work about relocation in free floating car sharing system and the simulator I use for the whole thesis. What’s more, I also review existing scientific literature that talks about the simulator modelling topics. Then comes to Chapter III that describes the simulator and dataset that used for the experiment in detail. Besides, I introduce both reactive and proactive relocation strategies. I present the results of a simulation campaign conducted for the city of Turin in Chapter IV. In the end, conclusion and future expectation are proposed in Chapter V.