Evolutionary Computing Group


Evolutionary Computing is a field of Computing Science where solution to problems are evolved and breed by means of Darwinian principles of natural selection. For each problem, the technique includes the art of

  • determining a computerized representation of the solution,
  • setting up a fitness function that quantitatively evaluates the quality of a candidate solution,
  • establishing a breed mechanism, that breeds new candidate solutions from the existing ones, based on their fitness value,
  • and determining which candidate solution individual has to die or live to see the next generation (in other words will be kept in the candidate solution pool or removed from it).

Evolutionary Computing is powerful because it:

  • has a great flexibility in representation, in other words it is not constrained, for example, to mathematical models/techniques.
  • is suitable for problems consisting of a very high number of parameters. EC solutions are usually of linear or quadratic complexity.
  • is suitable for heterogenous parameter types.
  • searches the solution spaces from hundreds of points.
  • encompases techniques to avoid local optima trapping.
  • mostly outperforms deterministic optimization techniques due to its stocastic nature.
  • is easy to implement multiobjectiveness like pareto-optimization.
  • is suitable for parallelization.

The EC group has solved numerous problems stemming from industry, military or daily life over its existence that exceeds a decade now. Furthermore, has contributed to science with direct contributions on the theoretical aspects of the field. Among these are:

  • Scheduling of final exams of a university
  • The post assignment problem of the armed forces
  • A new representation for combinatorial problems
  • A new technique in which the representation is dynamically rearranged
  • A new breeding mechanism for deceptive domains

Group members are offering graduate courses, MS and PhD thesis on the field.

The EC group is offering several services to industrial and institutional bodies:

  • Seminar activities on EC techniques
  • Consultancy that provides guidance in seeking EC solutions
  • Custom tailored EC solutions to specific problems

Selected Publications

  • İ. H. Toroslu, G. Üçoluk, Incremental Assignment Problem, Information Sciences, 177(6), p:1523, Elsevier, 2007
  • İ. H. Toroslu, Y. Arslanoğlu, Genetic algorithm for the personnel assignment problem with multiple objectives, Information Sciences, (177)3, p: 787, Elsevier, 2007
  • M. A. Bayır, İ H. Toroslu, A. Coşar, A Genetic algorithm for the multiple-query optimization problem, IEEE Transactions on Systems, Man, and Cybernetics, Part C, 37(1), p:147, 2007
  • E. Korkmaz, G. Üçoluk, A Controlled Genetic Programming Approach for the Deceptive Domain, IEEE Transactions on Systems, Man, and Cybernetics, Part B, 34(4), p:1730, 2004
  • G. Üçoluk, Genetic Algorithm Solution of the TSP Avoiding Special Crossover and Mutation, Intelligent Automation and Soft Computing, 3(8), TSI Press, 2002
  • G. Üçoluk, İ. H. Toroslu, A Genetic Algorithm Approach for Verification of the Syllable Based Text Compression Technique, Journal of Information Science, 23(5), Elsevier, 1997

Conference Proceedings

  • E. Korkmaz, G. Üçoluk, “Design and Usage of a New Benchmark Problem for Genetic Programming” Proceedings of ISCIS-2003, LNCS 2869, p:561, Springer Verlag, 2003
  • M. T. Yöndem, G. Üçoluk, “A Realistic Success Criterion for Discourse Segmentation” Proceedings of ISCIS-2003, LNCS 2869, p:592, Springer Verlag, 2003
  • O. T. Şehitoğlu, G. Üçoluk, “Gene Level Concurrency in Genetic Algorithms” Proceedings of ISCIS-2003, LNCS 2869, p:976, Springer Verlag, 2003
  • E. Korkmaz, G. Üçoluk, “Controlled Genetic Programming Search for Solving Deceptive Problems” Proceedings of the Genetic and Evolutionary Computation Conference, GECCO-2002, New York, 2002
  • O. T. Şehitoğlu, G. Üçoluk, “A Building Block Favoring Reordering Method for Gene Positions in Genetic Algorithms” Proceedings of the Genetic and Evolutionary Computation Conference, GECCO-2001. p:GA:571. San Francisco, 2001
  • E. Korkmaz, G. Üçoluk, “Genetic Programming for Grammer Induction” Proceedings of the Genetic and Evolutionary Computation Conference, GECCO-2001. p:GP:180. San Francisco, 2001
  • G. Üçoluk, “A Method for Chromosome Handling of r-Permutation of n-Element Set in Genetic Algorithms” in Proceedings of IEEE International Conference on Evolutionary Computation '97, p:80-85, Indianapolis (1997)