zaixing-sun / yifan-simulator-dynamic-workflow-scheduling-in-cloud-computing

[1] Y. Yang, G. Chen, H. Ma, S. Hartmann, and M. Zhang, “Enhancing generalization in genetic programming hyper-heuristics through mini-batch sampling strategies for dynamic workflow scheduling,” Inf. Sci. (Ny)., vol. 678, no. May, p. 120975, 2024, doi: 10.1016/j.ins.2024.120975.

Simulator - Dynamic Workflow Scheduling in Cloud Computing

Dynamic workflow scheduling (DWS) plays a crucial role in cloud computing. This simulator aims to make intelligent decisions in terms of workflow allocation and VM provisioning by VM selection rules (VMSRs). The objective is to minimize the total costs associated with VM rental fees and deadline violation penalties. Certainly, the objective function can also be customized according to needs.

Getting started

File Name	Description
MAIN-OneTimeSimulation.py	Evaluate a problem instance (with 30 workflows) using a randomly generated VM selection rule (in form of a GP tree)
MAIN-SGP.py	Using single-tree GPHH (SGP) to solve dynamic workflow scheduling problems, and individuals in each generation are evaluated on multiple problem instances. Training and test phases integrated into this file