Abstract
This paper focuses on the evaluation of a number of recently defined Evolution Strategy operators applied to a novel and challenging environment. The ES operates as the exploration component of a hybrid learning architecture for mobile robotics. In this on-line application both the population and number of generations must be small. Further, the objective function is multi-modal and dynamic, i.e. changing shape within and between generations. The results of our preliminary experiments indicate that derandomised mutation and intermediate recombination operators gave the best performance, especially with very small populations.
Preview
Unable to display preview. Download preview PDF.
References
D.A. Ackley. A Connectionist Machine for Genetic Hillclimbing. Kluwer Academic, Norwell, MA, 1987.
M.A. Arbib. Brains, Machines and Mathematics. McGraw-Hill, New York, 1964.
T. Bäck and H-P. Schwefel. An overview of evolutionary algorithms for parameter optimization. Evolutionary Computation, 1(1):1–23, 1993.
H-G. Beyer. Towards a theory of evolution strategies: On the benefits of sex — the (μ/μ, λ) theory. Evolutionary Computation, 3(1):81–111, 1995.
J. del R. Millán and C. Torras. A reinforcement connectionist approach to robot path finding in non-maze-like environments. Machine Learning, 8:363–395, 1992.
J.M. Fitzpatrick and J.J. Grefenstette. Genetic algorithms in noisy environments. Machine Learning, 3:101–120, 1988.
U. Hammel and T. Bäck. Evolution strategies on noisy functions: How to improve convergence properties. In Parallel Problem Solving from Nature, 3rd Workshop, PPSN3, pages 159–168, 1994.
F. Hoffmeister and T. Bäck. Genetic algorithms and evolutionary strategies: Similarities and differences. In Parallel Problem Solving from Nature, 1st Workshop, PPSN1, pages 455–469, 1990.
Long-Ji Lin. Reinforcement Learning for Robots Using Neural Networks. PhD thesis, Carnegie Mellon University, 1993.
A. Ostermeier, A. Gawelczyk, and N. Hansen. A derandomized approach to self-adaptation of evolution strategies. Evolutionary Computation, 2(4):369–380, 1995.
A.G. Pipe, T.C. Fogarty, and A. Winfield. Hybrid adaptive heuristic critic architectures for learning in mazes with continuous search spaces. In Parallel Problem Solving from Nature, 3rd Workshop, PPSN3, pages 482–490, 1994.
A.G. Pipe, T.C. Fogarty, and A. Winfield. A hybrid architecture for learning continuous environmental models in maze problems. In From Animals to Animats 3, Proceedings of the Third International Conference on Simulation of Adaptive Behaviour, pages 198–205, 1994.
G. Rudolph. Global optimisation by means of distributed evolution strategies. In Parallel Problem Solving from Nature, 1st Workshop, PPSN1, pages 209–213, 1990.
H-P. Schwefel. Numerical Optimization of Computer Models. Wiley, 1981.
H-P. Schwefel. Collective phenomena in evolutionary systems. In Preprints of the 31st Annual Meeting of the International Society for Generel Systems Research, pages 1025–1032, 1987.
H.A. Simon. The Sciences of the Artificial. MIT Press, Cambridge, MA, 1969.
R.S. Sutton. Temporal Credit Assignment in Reinforcement Learning. PhD thesis, University of Massachusetts, 1984.
R.S. Sutton. Learning to predict by the methods of temporal differences. Machine Learning, 3:9–44, 1988.
R.S. Sutton. First results with dyna, an integrated architecture for learning, planning and reacting. In W.T. Miller, R.S. Sutton, and P.J. Werbos, editors, Neural Networks for Control, pages 179–189. MIT Press, Cambridge, MA, 1990.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Sullivan, J.C.W., Pipe, A.G. (1996). Efficient Evolution Strategies for Exploration in mobile robotics. In: Fogarty, T.C. (eds) Evolutionary Computing. AISB EC 1996. Lecture Notes in Computer Science, vol 1143. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0032780
Download citation
DOI: https://doi.org/10.1007/BFb0032780
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-61749-5
Online ISBN: 978-3-540-70671-7
eBook Packages: Springer Book Archive