Two-stage stochastic programming based on the desirability function to optimize the performance of an internal-combustion engine

Abstract

The performance of an engine is one of the major concerns in the automotive industry. Increased emission restrictions on motor vehicles have led to the necessity of more accurate control on the engine performance. The objective of this study is to optimize the injection and ignition system in an automobile internal-combustion engine. In this paper the engine torque, the fuel consumption and the hydrocarbon emissions are considered as the main performance metrics of the engine, and a mathematical model is proposed on the basis of response surface methodology and two-stage stochastic programming. The proposed model encompasses desirability functions with a maximum–minimum operator to convert several responses into a single response and also considers the probabilistic pattern of regression coefficients by some random scenarios. To illustrate application of the proposed approach, it was applied in a real case in an automotive company. The results of this test represent a reasonable performance of the proposed approach.