Multi-Objective Optimization of HEV Fuel Economy and Emissions using Evolutionary Computation 2004-01-1153

How to match various components of the Hybrid Electric Vehicle (HEV) and manage the energy distribution is critical for HEV design and control considering that there are at least two sets of energy output systems, the fuel converter (engine or fuel cell) and the energy storage system (battery stack). It is an optimization issue. Much work has been done in this field but a significant portion optimizes the component sizes and control strategy parameters separately. What's more, most work concerns single objective optimization although more than one objective is more common and natural in HEV optimization. This paper uses an advanced soft computing technique, Non-dominated Sorting Genetic Algorithm (NSGA-II), one of the most efficient Multi-Objective Evolutionary Algorithms (MOEAs), to optimize a parallel HEV's fuel economy and emissions including HC, CO, and NOx emissions simultaneously, instead of converting them into a single objective. The task is to find the trade-off solutions, namely, Pareto-optimal solutions, among the four objectives. In addition, three component sizing parameters and four control strategy parameters are treated as variables all together. The obtained set of optimal solutions demonstrates great success either in converging to the real Pareto front or in providing sufficiently diverse solutions. It is concluded that the MOEA serves as a useful guide in the HEV developing process, especially in the earlier phases, since it can furnish the developers with a somewhat comprehensive preview of the problem.