DocumentCode :
3379268
Title :
Evolutionary potential timetables optimization by means of genetic and greedy algorithms
Author :
Arous, Najet ; Abdallah, Salah B. ; Ellouze, Noureddine
Author_Institution :
Dept. Electr., Ecole Nat. d´´Ingenieurs de Tunis, Tunisia
fYear :
1999
fDate :
1999
Firstpage :
24
Lastpage :
31
Abstract :
A highly constrained combinatorial problem like the timetable, can be solved by evolutionary methods (R. Piola, 1992). We describe an effective solution based on an alternation process involving the use of an appropriate genetic algorithm (GA) and a heuristic specific greedy algorithm. Both algorithms are conceived to respond to constraints and objectives imposed by the timetable problem. The proposed technique guarantees always to produce a feasible timetable in a reasonable time, computing by hard coding constraints which must not be broken. The goal of the presented is to prove that the alternate GA with specific exploration technique is a process that takes advantage of the global search of feasible solutions and specific technique efficiency in local solution optimization. Classic GA is not efficient when applied to the timetable problem, classified as NP hard. Alternating use of GA and the greedy algorithm is done so as to optimize resource distribution and thereafter to improve performance in terms of fitness values and time processing
Keywords :
computational complexity; genetic algorithms; scheduling; NP hard; alternate GA; alternation process; evolutionary methods; evolutionary potential timetable optimization; exploration technique; feasible timetable; fitness values; genetic algorithm; global search; greedy algorithms; hard coding constraints; heuristic specific greedy algorithm; highly constrained combinatorial problem; local solution optimizatio; local solution optimization; resource distribution; time processing; timetable problem; Constraint optimization; Dynamic programming; Electrical capacitance tomography; Genetic algorithms; Genetic programming; Greedy algorithms; Identity-based encryption; Random processes; Search methods; Simulated annealing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information Intelligence and Systems, 1999. Proceedings. 1999 International Conference on
Conference_Location :
Bethesda, MD
Print_ISBN :
0-7695-0446-9
Type :
conf
DOI :
10.1109/ICIIS.1999.810220
Filename :
810220
Link To Document :
بازگشت