In view of a joint optimization of cost and energy consumption in heterogeneous cloud systems (HCS), a swarm intelligence optimization algorithm is applied for task scheduling problems, with a reverse solution based white shark optimization algorithm (RS_WSO) to be proposed. As a metaheuristic algorithm, RS_WSO includes population initialization, calculation of reverse solutions, prey tracking and hunting. Experiments are carried out in two scientific workflows, involving epigenome (EP) and Gauss elimination (GE). The results show that RS_WSO algorithm is characterized with clear advantages in terms of cost saving and energy consumption compared with the current advanced meta-heuristic algorithm.