Optimization for Robust Design:
Integrating model-based systems engineering with multi-criteria decision-making support in a distributed framework
- Peter Fleming, Robin Purshouse, Visakan Kadirkamanathan, Ioannis Giagkiozis, Arun Chakrapani-Rao
- Funding sources: Engineering and Physical Sciences Research Council (EPSRC) and Jaguar Land Rover
This project forms part of the Programme for Simulation Innovation (PSi). PSi is a joint five-year research programme between Jaguar Land Rover Limited (JLR) and the Engineering and Physical Sciences Research Council (EPSRC). The PSi aim is to develop the capability of the virtual simulation industry in the UK and to give manufacturers like Jaguar Land Rover access to new, world-class simulation tools and processes.
The main objectives of our project are:
- To develop a design optimisation environment capable of producing robust designs at a variety of levels, ranging from enterprise level through to component level,
- To develop a framework that enables a network of decision support and optimisation nodes to operate asynchronously,
- To demonstrate these capabilities in two pilot case studies*, identified by Jaguar Land Rover as being of key strategic importance, and
- To develop an open source software tool that implements the framework and methods.
*An academic licence for Axisuite has been provided to enable us to undertake these case studies.
Liger Project - An Integrated Optimization Environment
- Peter J. Fleming, Ioannis Giagkiozis
- Funding source: Joint sponsorship by Ford Motors and The University of Sheffield.
New Horizons for Multi-Criteria Decision Making
The scheme involves internationally leading researchers from Brazil, Israel, Canada and Mexico and is supported under the International Research Staff Exchange Scheme (IRSES).
Platform for Rapid Application Development (RAISME)
System Architecture Design
Preference-Inspired Co-Evolutionary Algorithms (PICEAs)
The simultaneous optimisation of many objectives (say, in excess of 3), in order to obtain a full and satisfactory set of trade-off solutions to support a posteriori decision-making remains a challenging problem. The poor performance of Pareto-dominance based MOEAs (such as NSGA-II and SPEA2) is due to the fact that the proportion of non-dominated objective vectors in each MOEA population becomes very large as the number of objectives increases. As a result, not enough selection pressure can be generated toward the Pareto front The PhD project focus on a new class of algorithm called preferences-inspired co-evolutionary algorithms. This new approach co-evolves a family of decision makerâ€™s preferences with a population of candidate solutions during the search process. The candidate solutions are guided by the preferences to toward the true Pareto front. It is believed that this new approach is able to solve both bi-objective and many-objective problems.
Optimization Strategies for Power Management Systems
Energy resources onboard autonomous systems are limited. In recent years, rapid technology growth in autonomous systems has imposed a greater requirement of energy. Thus, an intelligent Power Management System (PMS) is required to maintain, or improve, the system capability. This project involves the development of an improved PMS. It is necessary that this adaptive and flexible PMS constructs optimal power plans while satisfying real-time requirements. The optimisation strategies developed for this PMS must be capable adapting to its dynamic environment, coping with any change in problem description, problem constraints, and problem objective