A study of permutation crossover operators on the traveling salesman problem
Proceedings of the Second International Conference on Genetic Algorithms on Genetic algorithms and their application
Minimum concave-cost network flow problems: applications, complexity, and algorithms
Annals of Operations Research
Revised-modified penalties for fixed charge transportation problems
Management Science
Bicriteria transportation problem by hybrid genetic algorithm
Proceedings of the 23rd international conference on on Computers and industrial engineering
Network random keys: a tree representation scheme for genetic and evolutionary algorithms
Evolutionary Computation
Evolution Strategies, Network Random Keys, and the One-Max Tree Problem
Proceedings of the Applications of Evolutionary Computing on EvoWorkshops 2002: EvoCOP, EvoIASP, EvoSTIM/EvoPLAN
Direct Representation and Variation Operators for the Fixed Charge Transportation Problem
PPSN VII Proceedings of the 7th International Conference on Parallel Problem Solving from Nature
Fitness Landscapes, Memetic Algorithms, and Greedy Operators for Graph Bipartitioning
Evolutionary Computation
Genetic algorithms, path relinking, and the flowshop sequencing problem
Evolutionary Computation
Edge sets: an effective evolutionary coding of spanning trees
IEEE Transactions on Evolutionary Computation
The edge-window-decoder representation for tree-based problems
IEEE Transactions on Evolutionary Computation
Expert Systems with Applications: An International Journal
A memetic algorithm for the quadratic multiple container packing problem
Applied Intelligence
Hi-index | 0.00 |
The fixed charge transportation problem (FCTP) is a classic challenge for combinatorial optimization; it is based on the well-known transportation problem (TP), and is one of the prime examples of an NP-complete variant of the TP, of general importance in a wide range of transportation network design problems. Many techniques have been applied to this problem, and the most effective so far (in terms of near-optimal results in reasonable time on large instances) are evolutionary algorithm based approaches. In particular, an EA proposed by Eckert and Gottlieb has produced the best performance so far on a set of specific benchmark instances. We introduce a new scheme, which has more general applicability, but which we test here on the FCTP. The proposed scheme applies an adaptive mutation process immediately following the evaluation of a phenotype. It thereby adapts automatically to learned information encoded in the chromosome. The underlying encoding approach is to encode an ordering of elements for interpretation by a constructive algorithm (such as with the Link and Node Biased encoding for spanning trees, and the Random Keys encoding which has been applied to both scheduling and graph problems), however the main adaptive process rewards links in such a way that genes effectively encode a measure of the number of times their associated link has appeared in selected solutions. Tests are done which compare our approach with Eckert and Gottlieb's results on benchmark FCTP instances, and other approaches.