Solving static optimal matching problem in heterogeneous processing with generalized stable marriage algorithms

Concepts and algorithms from a generalization of the stable marriage problem are used to optimally match machine features to task computational requirements in heterogeneous processing. Given a bilateral group-to-group linear preference order and a linear objective function, we can always find a one-to-one matching of machines and tasks in which no machine and task jointly prefer each other to their current assignments (stability). Among such matchings, there is one which is optimal with regard to the objective function. The algorithms to find it have polynomial-time complexity. Goals such as the majority assignment Pareto efficiency and static load-balancing are achieved simultaneously.<
>