Reinforced explorit on optimizing vehicle powertrains

Victor Parque, Masakazu Kobayashi, Masatake Higashi

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

10 Citations (Scopus)


How to build optimal vehicular powertrains? We study this question and propose an algorithm inspired by a domain-general design process. The basic idea is to interplay co-biasingly between the local approximations of discrete design and the global refinements of continuous parameters. The proposed method was evaluated to design powertrains of four types of vehicles: Series Hybrid Electric Vehicle(SHEV), Parallel Hybrid Electric Vehicle(PHEV), Fuel Cell(FC) and Electric Vehicle(EV). Simulation results show noticeable improvements on mileage per gas emissions over different study cases. To our knowledge, this is the first study aiming at designing vehicle powertrains considering the holistic point of view.

Original languageEnglish
Title of host publicationNeural Information Processing - 20th International Conference, ICONIP 2013, Proceedings
Number of pages8
EditionPART 2
Publication statusPublished - 2013 Dec 1
Externally publishedYes
Event20th International Conference on Neural Information Processing, ICONIP 2013 - Daegu, Korea, Republic of
Duration: 2013 Nov 32013 Nov 7

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume8227 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other20th International Conference on Neural Information Processing, ICONIP 2013
Country/TerritoryKorea, Republic of


  • Electric vehicle
  • Explorit
  • Fuel cell
  • Parallel hybrid electricvehicle
  • Reinforcement learning
  • Series hybrid electric vehicle

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Reinforced explorit on optimizing vehicle powertrains'. Together they form a unique fingerprint.

Cite this