Saad A. Parallel realization of genetic algorithms for scheduling tasks assigned on permutation.

Dissertation studies models of arrangement tasks assigned on permutations and develops methods of their optimization. Most of the time-consuming tasks of scheduling theory can be reduced to tasks assigned on permutations. As a rule these tasks are NP-complete. To solve them a modification of genetic algorithms is proposed in the dissertation as well as a method based on a combination of branches and boundaries method and genetic algorithm is proposed to get more exact solutions with a given accuracy. To increase performance of the mentioned methods parallel versions and realization on cluster computer systems are proposed. A package of applications for educational schedules compiling is created on the basis of received results