Increasing debates over a gasoline independent future and the reduction of greenhouse gas (GHG) emissions has led to a surge in plug-in hybrid electric vehicles (PHEVs) being developed around the world. Due to the limited all-electric range of PHEVs, a daytime PHEV charging infrastructure will be required for most PHEVs’ daily usage. This dissertation, for the first time, presents a mixed integer mathematical programming model to solve the PHEV charging infrastructure planning (PCIP) problem. Our case study, based on the Oak Ridge National Laboratory (ORNL) campus, produced encouraging results, indicates the viability of the modeling approach and substantiates the importance of considering both employee convenience and appropriate grid connections in the PCIP problem. Unfortunately, the classical optimization methods do not scale up well to larger practical problems. In order to effectively and efficiently attack larger PCIP problems, we develop a new MASTS based TS algorithm, PCIP-TS to solve the PCIP. The results from computational experiments for the ORNL campus problem establish the dominant supremacy of the PCIP-TS method both in terms of solution quality and computational time. Additional experiments with simulated data representative of a problem that might be faced by a small city show that PCIP-TS outperforms CPLEX based optimization. Once the charging infrastructure is in place, the immediate problem is to judiciously manage this system on a daily basis. This thesis formally develops a mixed integer linear program to solve the daily the energy management problem (DEM) faced by an organization and presented results of a case study performed for ORNL campus. The results from our case study, based on the Oak Ridge National Laboratory (ORNL) campus, are encouraging and substantiate the importance of controlled PHEV fleet charging and realizing V2G capabilities as opposed to uncontrolled charging methods. Although optimal solutions are obtained, the solver requires practically unacceptable computational times for larger problems. Hence, we develop a new MASTS based TS algorithm, DEM-TS, for the DEM models. Results for ORNL campus data set prove the dominant computational efficiency of the DEM-TS. For the simulated extended sized problems that resemble the complexity of a problem faced by a small city, the results prove that DEM-T not only achieves optimality, but also produces sets of multiple alternate optimal solutions. These could be very helpful in practical settings when alternate solutions are necessary because some solutions may not be deployable due to unforeseen circumstances. / text