Real World Optimisation with Life-Long Learning (ROLL)

01/01/2013 - 31/12/2015

project image
This project aims to improve the current state of the art in developing optimisation tools which are relevant and acceptable to industry.

This will be achieved by addressing industrial current concerns regarding the ability of academic optimisation techniques to deal effectively with highly constrained real-world problems and the cost and expertise required to develop and maintain these tools. The project addresses these concerns through the following objectives:

1. Development of a novel hyper-heuristic optimisation system which exhibits lifelong learning; the system will maintain and exploit a database of knowledge to produce fast high-quality solutions to problems while simultaneously autonomously adapting to dynamically changing problem characteristics in order to improve its performance over time and react to changes in its problem solving environment.

2. Demonstration that the proposed life-long learning system is more efficient and effective at rapidly producing high-quality solutions to  real-world practical problems than current optimisation approaches, producing savings from a financial perspective but simultaneously addressing environmental concerns regarding sustainability of operation and reduction of carbon emissions.

3. Engaging with end-users to develop an information database of problem-solving knowledge in a range of practical domains as a platform to drive advances in optimisation algorithms. Incorporating a suite of problem generators, problem libraries and heuristics, the design of the platform will be informed by real-world problems, encapsulating detailed practical constraints coupled with performance criteria, both of which will be specified by industrial experts.

4. Facilitating the uptake and use of optimisation tools within industry by demonstrating that the optimisation tool proposed is transferable across problem domains, does not require expert knowledge to develop or tune, and is therefore cheap to implement,  has considerably reduced maintenance costs due to its ability to autonomously improve over time and accounts for the highly constrained and often idiosyncratic nature of real-world problems.
Real World Optimisation with Life-Long Learning (ROLL) is a Research Councils project funded by EPSRC. Carried out in collaboration with Optrak and others. For further information please refer to
[Read More]

Related Projects

next prev

Areas of Expertise link icon

  • Bio-inspired Computing
    The Bio-Inspired Algorithms group within the Centre for Algorithms, Visualisation and Evolving Systems is a large and thriving group with interests in nature-inspired computing that include Evolutionary Computing, Hyper-Heuristics, Artificial Immune Systems and Swarm Intelligence.
next prev

Project Team

Emma Hart
Director of CEC
+44 131 455 2783
Kevin Sim
+44 131 455 2497

Associated Publications

Hart, E., Sim, K. (2016). A Hyper-Heuristic Ensemble Method for Static Job-shop Scheduling. Evolutionary Computation, (pre-print, accepted for publication May 2016), (), .

Sim, K., Hart, E. (2016). A Combined Generative and Selective Hyper-heuristic for the Vehicle Routing Problem. In: (Ed.) Proceedings go GECCO 2016, , () ( ed.). (pp. ). : . ACM.

Hart, E., Sim, K., Paechter, B. (2015). A Lifelong Learning Hyper-heuristic Method for Bin Packing. Evolutionary Computation, 23, (1), 37-67.

Hart, E., Sim, K. (2015, ). A Research Agenda for Metaheuristic Standardization. Paper presented at 11th Metaheuristics International Conference, Agadir.

Sim, K., Hart, E., Urquhart, N., Pigden, T. (2015). A new rich vehicle routing problem model and benchmark resource. In: (Ed.) Proceedings of the The 11th edition of the International Conference on Evolutionary and Deterministic Methods for Design, Optimization and Control with Applications to Industrial and Societal Problems (EUROGEN 2015), , () ( ed.). (pp. ). : . .

Sim, K., Hart, E. (2015). A Novel Heuristic Generator for JSSP Using a Tree-Based Representation of Dispatching Rules. In: (Ed.) Proceedings GECCO 2015 Companion, , () ( ed.). (pp. ). : . ACM Association for Computing Machinery.

See all publications