Efficient approximation algorithms for scheduling moldable tasks

Xiaohu Wu, Patrick Loiseau

Abstract: Highlights•We study scheduling independent moldable tasks under the proposed (δj,kj)-monotonic model, whose generality is between the classic monotonic and linear-speedup models.•An efficient algorithm for makespan minimization is proposed, achieving an approximation ratio close to 1.333 under mild assumptions.•Additionally, an efficient algorithm for throughput maximization with a deadline is proposed, achieving an approximation ratio close to 0.75 under mild assumptions.