ABSTRACT
Multiple assessment directed novelty search (MADNS), introduced by the authors in [20], is an extension to the novelty search algorithm which exploits the observation that populations optimised for phenotypic novelty may contain solutions to multiple independent and conflicting objectives. It has been shown that, through the application of MADNS, an evolutionary trajectory may be simultaneously directed towards multiple conflicting objectives. Previous results from a series of simulated maze navigation experiments have shown that MADNS may significantly outperform novelty search in domains where the potential for phenotypic exploration is high [20]. In this paper we further explore the MADNS algorithm, assessing the effect upon the diversity and performance of the population as the phenotypic land- scape increases. A series of experiments in domains with multiple conflicting objectives and expanding areas of irrelevant space show that the relative performance gain of MADNS increases alongside the potential for exploration. We conclude that, as the potential for exploration within a domain expands, the importance of directing novelty becomes ever more necessary.
- G. Cuccu and F. Gomez. When novelty is not enough. In Applications of Evolutionary Computation, pages 234--243. Springer, 2011. Google ScholarDigital Library
- K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan. A fast and elitist multiobjective genetic algorithm: Nsga-ii. Evolutionary Computation, IEEE Transactions on, 6(2):182--197, 2002. Google ScholarDigital Library
- S. Doncieux and J.-B. Mouret. Dynamic behavioral diversity. In Proceedings of the 14th International Conference on Genetic and Evolutionary Computation Conference, pages 1453--1454. ACM, 2012. Google ScholarDigital Library
- M. G. Epitropakis, X. Li, and E. K. Burke. A dynamic archive niching differential evolution algorithm for multimodal optimization. In Evolutionary Computation (CEC), 2013 IEEE Congress on, pages 79--86. IEEE, 2013.Google ScholarCross Ref
- J. Gomes, P. Mariano, and A. L. Christensen. Devising effective novelty search algorithms: A comprehensive empirical study. In Proceedings of the 2015 on Genetic and Evolutionary Computation Conference, pages 943--950. ACM, 2015. Google ScholarDigital Library
- J. Gomes, P. Urbano, and A. L. Christensen. Progressive minimal criteria novelty search. In Advances in Artificial Intelligence--IBERAMIA 2012, pages 281--290. Springer, 2012.Google ScholarCross Ref
- F. J. Gomez. Sustaining diversity using behavioral information distance. In Proceedings of the 11th Annual conference on Genetic and evolutionary computation, pages 113--120. ACM, 2009. Google ScholarDigital Library
- S. Kistemaker and S. Whiteson. Critical factors in the performance of novelty search. In Proceedings of the 13th annual conference on Genetic and evolutionary computation, pages 965--972. ACM, 2011. Google ScholarDigital Library
- J. Lehman and K. O. Stanley. Exploiting open-endedness to solve problems through the search for novelty. In ALIFE, pages 329--336, 2008.Google Scholar
- J. Lehman and K. O. Stanley. Efficiently evolving programs through the search for novelty. In Proceedings of the 12th Annual Conference on Genetic and Evolutionary Computation, GECCO '10, pages 837--844. ACM, 2010. Google ScholarDigital Library
- J. Lehman and K. O. Stanley. Revising the evolutionary computation abstraction: minimal criteria novelty search. In Proceedings of the 12th annual conference on Genetic and evolutionary computation, pages 103--110. ACM, 2010. Google ScholarDigital Library
- J. Lehman and K. O. Stanley. Abandoning objectives: Evolution through the search for novelty alone. Evolutionary computation, 19(2):189--223, 2011. Google ScholarDigital Library
- J. Lehman and K. O. Stanley. Evolving a diversity of virtual creatures through novelty search and local competition. In Proceedings of the 13th annual conference on Genetic and evolutionary computation, pages 211--218. ACM, 2011. Google ScholarDigital Library
- A. Liapis, G. N. Yannakakis, and J. Togelius. Enhancements to constrained novelty search: Two-population novelty search for generating game content. In Proceedings of the 15th annual conference on Genetic and evolutionary computation, pages 343--350. ACM, 2013. Google ScholarDigital Library
- J.-B. Mouret. Novelty-based multiobjectivization. In New horizons in evolutionary robotics, pages 139--154. Springer, 2011.Google ScholarCross Ref
- J.-B. Mouret and S. Doncieux. Encouraging behavioral diversity in evolutionary robotics: An empirical study. Evolutionary computation, 20(1):91--133, 2012. Google ScholarDigital Library
- M. Preuss, A. Liapis, and J. Togelius. Searching for good and diverse game levels. In Computational Intelligence and Games (CIG), 2014 IEEE Conference on, pages 1--8. IEEE, 2014.Google ScholarCross Ref
- S. Risi, S. D. Vanderbleek, C. E. Hughes, and K. O. Stanley. How novelty search escapes the deceptive trap of learning to learn. In Proceedings of the 11th Annual conference on Genetic and evolutionary computation, pages 153--160. ACM, 2009. Google ScholarDigital Library
- D. Smith, L. Tokarchuk, and C. Fernando. Evolving diverse strategies through combined phenotypic novelty and objective function search. In Applications of Evolutionary Computation, pages 344--354. Springer, 2015.Google ScholarCross Ref
- D. Smith, L. Tokarchuk, and G. Wiggins. Exploring conflicting objectives with madns: Multiple assessment directed novelty search. In Proceedings of the Companion Publication of the 2016 Annual Conference on Genetic and Evolutionary Computation, GECCO Companion '16. ACM, 2016. Google ScholarDigital Library
- K. O. Stanley and R. Miikkulainen. Evolving neural networks through augmenting topologies. Evolutionary computation, 10(2):99--127, 2002. Google ScholarDigital Library
Index Terms
- Harnessing Phenotypic Diversity towards Multiple Independent Objectives
Recommendations
Exploring Conflicting Objectives with MADNS: Multiple Assessment Directed Novelty Search
GECCO '16 Companion: Proceedings of the 2016 on Genetic and Evolutionary Computation Conference CompanionNovelty search is an evolutionary approach which promotes phenotypic diversity in a population. Novelty search has been successfully applied to a wide range of domains and a number of variants have been proposed. Here we introduce Multiple Assessment ...
Evolution of heterogeneous multirobot systems through behavioural diversity
AAMAS '14: Proceedings of the 2014 international conference on Autonomous agents and multi-agent systemsHeterogeneity is present in many collective systems found in nature and considered fundamental for effective task execution in several complex, real-world scenarios. Evolutionary computation has the potential to automate the design of multirobot systems,...
Effective diversity maintenance in deceptive domains
GECCO '13: Proceedings of the 15th annual conference on Genetic and evolutionary computationDiversity maintenance techniques in evolutionary computation are designed to mitigate the problem of deceptive local optima by encouraging exploration. However, as problems become more difficult, the heuristic of fitness may become increasingly ...
Comments