Reliable Scheduling of Precedence-Const rained Tasks Using a Genetic Algorithm

摘要

This paper investigates the problem of reliable matching and scheduling of an application, which is composed of tasks with precedence constraints, in a heterogeneous distributed computing system. The reliable matching and scheduling problem is posed as finding a task assignment to minimize the probability of failure of the application. A cost function which defines this probability under a given task assignment is derived. Because the time complexity of the reliable matching and scheduling problem is exponential, to find optimal and subopti-mal solutions efficiently, a novel genetic algorithm is devised. The simulation results are provided to confirm the performance of the proposed genetic al-gorithm.