An Ant Colony Algorithm for Improving Energy Efficiency of Road Vehicles

Energies

Alberto V. Donati¹  Christian Thiel¹  Jette Krause¹  Ben White²  Nikolas Hill³ 

[1] Joint Research Centre of the European Commission, Via Fermi 2749, 21027 Ispra (VA), Italy;Ricardo Energy and Environment, 30 Eastbourne Terrace, Paddington, London W2 6LA, UK;Ricardo Energy and Environment, Gemini Building, Fermi Avenue, Harwell, Oxon OX11 0QR, UK;

关键词: CO2 reduction;    multi-objective combinatorial optimization;    meta-heuristics;    ant colony optimization;   

DOI  :  10.3390/en13112850

来源: DOAJ

【 摘 要 】

The number and interdependency of vehicle CO₂ reduction technologies, which can be employed to reduce greenhouse emissions for regulatory compliance in the European Union and other countries, has increasingly grown in the recent years. This paper proposes a method to optimally combine these technologies on cars or other road vehicles to improve their energy efficiency. The methodological difficulty is in the fact that these technologies have incompatibilities between them. Moreover, two conflicting objective functions are considered and have to be optimized to obtain Pareto optimal solutions: the CO₂ reduction versus costs. For this NP-complete combinatorial problem, a method based on a metaheuristic with Ant Colony Optimization (ACO) combined with a Local Search (LS) algorithm is proposed and generalized as the Technology Packaging Problem (TPP). It consists in finding, from a given set of technologies (each with a specific cost and CO₂ reduction potential), among all their possible combinations, the Pareto front composed by those configurations having the minimal total costs and maximum total CO₂ reduction. We compare the performance of the proposed method with a Genetic Algorithm (GA) showing the improvements achieved. Thanks to the increased computational efficiency, this technique has been deployed to solve thousands of optimization instances generated by the availability of these technologies by year, type of powertrain, segment, drive cycle, cost type and scenario (i.e., more or less optimistic technology cost for projected data) and inclusion of off-cycle technologies. The total combinations of all these parameters give rise to thousands of distinct instances to be solved and optimized. Computational tests are also presented to show the effectiveness of this new approach. The outputs have been used as basis to assess the costs of complying with different levels of new vehicle CO₂ standards, from the perspective of different manufacturer types as well as vehicle users in Europe.

【 授权许可】

Unknown