Reduction of the driving cost for capacitor-battery combined HEV by using genetic network programming

The purpose of this study is reduction of both fuel consumption and battery deterioration by the suboptimal control for hybrid electric vehicles (HEV) which can be adapted for various driving situations and drivers as part of the emergent control system. From the rising price of fossil fuels and growing concerns on the environmental protection, HEVs are recently attracting attention. The HEV control system requires reduction of fuel consumption and battery deterioration, as well as adaptivity to dynamic environment and various drivers. However no conventional methods have taken this in account. Consequently in this paper, we use genetic network programming (GNP) as a method to find a suboptimal control of HEVs for various situations. GNP is one of the evolutionary computing methods, which has a network structure as its solution and is effective to dynamical problem because it implicitly depends on past information. We validated the proposed method by applying it to the driving simulation. Comparison to the conventional method proved the advantage of the new proposed method. We introduced new evaluation criterion "driving cost [¥/km]" in order to consider all the cost needed to drive; not only the fuel cost nor the battery cost. In addition, we validated the adaptivity to various driving situation and various drivers.