Multi-criterion Optimisation of Distributed System Performance by Evolutionary Task Assignments


Abstract

A careful task assignment in a distributed computer system may reduce the workload of a bottleneck computer. It can decrease the cost of computation if the computer sort selection is applied with the task assignment. The total system performance is another measure that can be minimised by the computer sort selection with the task assignment. An extended problem of task allocation can be formulated as a multiobjective combinatorial optimisation question, which is solved by an adaptive evolutionary algorithm. It is applied by finding the subset of Pareto-optimal solutions. Then, a module scheduling is used to maximise the probability of completing tasks with timing correctness.