| Summary: | 博士 === 國立臺北科技大學 === 電子工程系 === 107 === In metropolitan areas, drivers share their vehicles with people during daily commutes using a carpool process. For the procedure of sharing empty seats, we need to consider increased ridership and driving distances incurred by carpool detours resulting from matching passengers to drivers, maximizing the number of simultaneous matches, as well as the driver/passenger participating in carpooling specifies an available departure time for the following itinerary and he/she could be delayed for some reason. In accordance with these goals, we first defined a multiobjective carpool service problem with time windows (MOCSPTW) by considering four optimized objectives. This thesis developed a Set-based Simulated Binary Operation (SSBO) which improving the individual representation and genetic operation in evolutionary algorithm. Subsequently, a coevolutionary algorithm for two solution sets, population and archive, using objective-wise local search and set-based simulated binary operation (COEA-LS-SSBO) was proposed. In the experimental result, the proposed SSBO can provide better driver-passenger matching results than can the binary-coded and set-based non-dominated sorting genetic algorithms. The results of quantitative comparison and the objectives visualization showed that the proposed coevolutionary algorithm can obtain better Pareto-optimal solutions than a fast, non-dominated sorting genetic algorithm in terms of convergence and diversity.
|
|---|
