Feasibility Assessment and Operating Policy Optimization of Automotive Powertrains With Uncertainties Using Game Theory

Abstract

The paper describes a game-theoretic framework and a computational algorithm for feasibility evaluation of automotive powertrains with storage elements in terms of fuel economy and emissions performance. The game-theoretic framework allows to handle various time-dependent uncertainties, including uncertainties in the drive cycle. In particular, an important issue in the prior approaches to this problem, the drive cycle dependence of the optimal policies, is alleviated. Within the basic framework, it is also possible to generate implementable operating policies that specify powertrain actuator settings as functions of engine operating conditions and states of the storage elements. We illustrate the procedure using an example powertrain with a Direct Injection Stratified Charge engine and an aftertreatment system consisting of a Three Way Catalyst and a Lean NOx Trap.