Adaptive engine optimisation using NSGA-II and MODA based on a sub-structured artificial neural network

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Colleges, School and Institutes

External organisations

  • National Nanotechnology Laboratory (NNL) of CNR-INFM, Via per Arnesano 73100 Lecce, Italy


The trend in the automotive industry is towards electric vehicles (EV), however, the industry will depend on gasoline engines for many years to come. There is also increased demand for the reduction of greenhouse gases. This work develops an adaptive model-based optimal control algorithm based on Sub-Structured Neural Network (SSANN), Multi-Objective Genetic Algorithms (GA), Multi-Objective Dragonfly Algorithm (MODA) and a fuzzy based inference system. The SSANN based on an individual engine speed are combined into a SSANN, whose output is connected to a fuzzy based inference system which extrapolates between trained engine speeds. The SSANN outputs are then used to evaluate the objective function of the optimisation process, which is performed using GA and MODA. The SSANN is retrained if the error is out of defined limit, which allows the system to adapt with engine ageing. The purpose of this work is to contribute to online engine calibration and control to improve engine performance and reduce greenhouse gas emissions.


Original languageEnglish
Title of host publicationICAC 2017 - 2017 23rd IEEE International Conference on Automation and Computing
Subtitle of host publicationAddressing Global Challenges through Automation and Computing
Publication statusPublished - 23 Oct 2017
Event23rd IEEE International Conference on Automation and Computing, ICAC 2017 - Huddersfield, United Kingdom
Duration: 7 Sep 20178 Sep 2017


Conference23rd IEEE International Conference on Automation and Computing, ICAC 2017
CountryUnited Kingdom