Adaptation in natural and artificial systems
Adaptation in natural and artificial systems
Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
A Genetic Algorithm for Survivable Network Design
Proceedings of the 5th International Conference on Genetic Algorithms
Shall We Repair? Genetic AlgorithmsCombinatorial Optimizationand Feasibility Constraints
Proceedings of the 5th International Conference on Genetic Algorithms
Design of Stacked Self-Healing Rings Using a Genetic Algorithm
Journal of Heuristics
| Hi-index | 0.00 |
Suppose that items of equipment are to be added to a supply station (e.g., new switch modules are to be added to a telecommunications switch)over time to meet growing demand requirements. Both supply and demand havemultiple components: an item of equipment supplies different amounts ofseveral resources, and demand may be expressed in terms of the vector ofresources required. There are several different types of equipment to chooseamong, each type supplying known amounts of each resource per unit ofequipment. The supply station is organized into bays, shelves, or othercapacitated “containers” so that when the cumulative amount ofequipment added exceeds the holding capacity of the installed containers,new containers must be added, creating a relatively large jump in cumulativecosts. Thus, it is desirable to sequentially “pack” items ofequipment into the available containers, by choosing which types ofequipment to install when, so as to minimize the total cost of coveringdemand in each period. We discuss an instance of this problem arising fromwireless telephony and describe the performance of a conventionalbranch-and-bound optimization algorithm for solving it. The branch-and-boundapproach works well on small instances of the problem, and has been usedsuccessfully in practical planning. However, it can take CPU-days to run,thus preventing development of a useful interactive planning tool.Therefore, we introduce a novel “seed, repair, and replace”genetic algorithm (SRR-GA) for solving dynamic packing problems of thistype. We contrast its performance with the branch-and-bound algorithm‘s onboth hand-generated and randomly-created dynamic packing problems, findingthat the SRR-GA is two to three orders of magnitude faster and producessolutions of equal or better quality on practical problems. Variations ofthe dynamic packing problem and of the SRR-GA for solving it are mentioned,and the paper concludes by suggesting other potential applications of theSRR-GA to hard combinatorial optimization problems.