Overview:
Optimization problems in which the evaluation of solutions is expensive arise in a variety of contexts. The reasons for the high cost of evaluation and their effect on how many evaluations/generations can be afforded differ widely from one problem to another, as the following three examples may illustrate. (i) When evolving controllers for a simulated collective of robots, the fidelity of the physics simulator, the noise/stochasticity in the system, and the desire to obtain robots that are robust to rare events may all play a part in making simulation times very long. (ii) When evolving a novel protein for a specific binding target by synthesis of proteins in vitro and their subsequent screening, thousands of proteins may be synthesised in parallel but each further "generation" will take another 12 hours to process and will also have financial implications. (iii) When evolving a basic conceptual design for a new building, an architect evaluating the designs will suffer fatigue after several hours and will eventually have to stop.
Scope:
This special session invites contributions in all aspects of applying evolutionary computation to the optimization of expensive optimization problems, to include the following areas.
- The use of meta-models and their integration in evolutionary algorithms.
- Evolutionary algorithms using fitness inheritance.
- Theory of on-line and off-line learning for expensive optimization.
- Validation techniques for approximate landscape models.
- Empirical comparison of meta-modelling approaches.
- Statistical performance analyses of surrogate-assisted EAs, to include worst-case performance estimation.
- Off-line learning and landscape state machines.
- Incorporating local search or approximated gradient descent into surrogate-assisted EAs.
- Test functions designed specifically to model particular expensive optimization problems.
- Expensive optimization problems with noise and/or constraints.
- Combinatorial and mixed integer expensive optimization problems.
- Effects of the time taken to perform one evaluation on algorithm choice. Effects of population size limit on algorithm choice.
- Multiobjective methods for expensive optimization problems.
- Using landscape statistics (e.g. epistasis variance, correlation length, estimates of number of optima) to adapt search strategies on-line or off-line.
- Design-of-experiments initialization schemes/principles for use in EAs.
- Operators, selection schemes, niching methods and adaptation schemes for use with EAs using surrogate modeling.
- Intensification/diversification issues in expensive optimization problems.
Paper submission:
Manuscripts should be prepared according to the standard format of regular papers specified in CEC 2006 and be restricted to a maximum of 8 pages. (Note: the page restriction was formerly 6 pages). Instructions for preparing the paper are provided here. Paper submission is strictly only PDF format and online through the regular CEC2006 submission website. Please select the topic coding "Zd" as the "Main Research Topic" when you submit. Special session papers will be treated in the same way as regular papers and included in the conference proceedings.
Contact:
The special session is co-organized by
- Joshua Knowles (University of Manchester, UK) [Homepage]
- Yew Soon Ong (Nanyang Technological University, Singapore) [Homepage] ,
Scientific program committee members:
- Jürgen Branke, Institute AIFB, University of Karlsruhe, Germany
- David Corne, Heriot-Watt University, UK
- Marco Farina, AST - Advanced System Technology, STMicroelectronics, Italy
- Kyriakos Giannakoglou, National Technical University of Athens, Greece
- Evan Hughes, Cranfield University, UK
- Yaochu Jin, Honda Research Institute Europe (HRI-EU), Offenbach/Main, Germany
- Andy Keane, University of Southampton, UK
- Khaled Rasheed, University of Georgia, GA, US
- Bernhard Sendhoff, Honda Research Institute Europe (HRI-EU), Offenbach/Main, Germany
- Wei Shyy, Department of Aerospace Engineering, University of Michigan, US
- Lars Willmes, NuTech Solutions GmbH, Dortmund, Germany
Sponsors:
CEC 2006 is sponsored and supported by The Institute of Electrical and Electronic Engineers, The IEEE Computational Intelligence Society, the International Neural Network Society, the Institute of Electrical Engineers and the Evolutionary Programming Society.