Solutions to Parallel and Distributed Computing Problems: Lessons from Biological Sciences

5.5 SUMMARY

The significance of the work produced here is the development of the StrucGene framework, which involves the use of a structured model to represent the problem, followed by the use of a genetic algorithm to obtain satisfactory solutions to the scheduling problem. Another attractive feature of genetic algorithms is the fact that such structures have a highly parallelizable nature, which can be exploited to create faster versions of the developed algorithms [3].

Scheduling problems involving complex precedence relations, processing costs, and ITC costs can be represented systematically and easily through the use of matrices. A genetic algorithm, which is a robust optimization tool, can be used to generate assignments that can produce near-optimal solutions in a rapid fashion. The StrucGene framework provides the flexibility of extending the scheduling problem to an arbitrary number of processors with little or no effort.