ABSTRACT
In this paper we consider the problem of extended finite-state machines induction. The input data for this problem is a set of tests. Each test consists of two sequences - an input sequence and a corresponding output sequence. We present a new method of Extended Finite-State Machines (EFSM) induction based on an Ant Colony Optimization algorithm (ACO) and a new meaningful test-based crossover operator for EFSMs. New algorithms are compared with a genetic algorithm (GA) using a traditional crossover, a (1+1) evolutionary strategy and a random mutation hill climber. This comparison shows that the use of test-based crossover greatly improves performance of GA. GA on average also significantly outperforms the hill climber and evolutionary strategy. ACO outperforms GA, and the difference between average performance of ACO and GA hybridized with hill climber is insignificant.
- Clark J., et al. Reformulating Software Engineering as a Search Problem. IEEE Proceedings-Software, vol. 150, 3, 2003, pp. 161--175.Google Scholar
- Harman M., Mansouri A. and Zhang Y. Search-Based Software Engineering: A Comprehensive Analysis and Review of Trends, Techniques, and Applications, tech. report TR- 09-03, Dept. of Computer Science, King's College London, 2009.Google Scholar
- Harman M. Software Engineering Meets Evolutionary Computation // Computer. 2011. Vol. 44, 11, pp. 31 -- 39. Google ScholarDigital Library
- Polikarpova, N. and Shalyto, A. 2009. Automata-based programming. Piter (in Russian).Google Scholar
- Dorigo, M. 1992. Optimization, Learning and Natural Algorithms. PhD thesis. Polytechnico di Milano, Italy.Google Scholar
- Lucas, S. and Reynolds, J. Learning DFA: Evolution versus Evidence Driven State Merging. The 2003 Congress on Evolutionary Computation (CEC'03). Vol. 1, 351--348.Google Scholar
- Lucas, S. Evolving Finite-State Transducers: Some Initial Explorations. Lecture Notes in Computer Science. Springer Berlin / Heidelberg. Volume 2610/2003, pp. 241--257.Google Scholar
- Hamming, R. Error detecting and error correcting codes. Bell System Technical Journal 29 (2), pp. 147--160.Google ScholarCross Ref
- Levenshtein, V. Binary codes capable of correcting deletions, insertions, and reversals. Soviet Physics Doklady 10: 707--10. 1966.Google Scholar
- Johnson, C. Genetic Programming with Fitness based on Model Checking. Lecture Notes in Computer Science. Springer Berlin / Heidelberg, 2007. Volume 4445/2007, pp. 114--124. Google ScholarDigital Library
- Lucas, S. and Reynolds, J. Learning Deterministic Finite Automata with a Smart State Labeling Algorithm. IEEE Transactions on Pattern Analysis and Machine Intelligence. Vol. 27, 7, 2005, pp. 1063--1074. Google ScholarDigital Library
- Zaikin, A. Development of finite-state machines induction methods with simulated annealing for the "Competition for resources" game. Nauchno-technicheskij vestnik SPbGU ITMO (in Russian). 2011. 2, pp. 49 -- 54.Google Scholar
- Tsarev, F. 2010. Method of finite-state machine induction from tests with genetic programming. Information and Control Systems (Informatsionno-upravljayushiye sistemy, in Russian), no. 5, pp. 31--36. http://is.ifmo.ru/works/_zarev.pdfGoogle Scholar
- Tsarev, F. and Egorov, K. Finite-state machine induction using genetic algorithm based on testing and model checking. 2011. In Proceedings of the 2011 GECCO Conference Companion on Genetic and Evolutionary Computation (GECCO'11). NY. ACM. 2011. pp. 759--762. Google ScholarDigital Library
- Alba, E. and Chicano, F. 2007. ACOhg: dealing with huge graphs. In Proceedings of the 9th annual conference on Genetic and evolutionary computing (GECCO'07). ACM, NY, USA, pp. 10--17. Google ScholarDigital Library
Recommendations
Binary-coding-based ant colony optimization and its convergence
AbstractAnt colony optimization (ACO for short) is a meta-heuristics for hard combinatorial optimization problems. It is a population-based approach that uses exploitation of positive feedback as well as greedy search. In this paper, genetic algorithm's (...
Cooperative ant colony-genetic algorithm based on spark
By taking full advantages of both the map and reduce function for the MapReduce parallel framework and the memory computation for the Spark platform, this paper designs and implements the algorithms for solving the traveling salesman problem based on ...
A new crossover mechanism for genetic algorithms with variable-length chromosomes for path optimization problems
A new crossover mechanism containing two operators is proposed.The new mechanism produces more cross point pairs than same point crossover.Our novel genetic algorithm has a desirable performance. Genetic Algorithm (GA) has found wide application in path ...
Comments