Reinforcement learning based an Integrated Approach for Uncertainty Scheduling in Adaptive Environments using MARL

Scheduling is a universal theme being conferred in technological areas like computing and strategic areas like operational management. The core idea behind scheduling is the distribution of shared resources across time for competitive tasks. Optimization, efficiency, productivity and performance are the major metrics evaluated in scheduling. Effective scheduling under uncertainty is tricky and unpredictable and its an interesting area to study. Environmental uncertainty is a challenging extent that effect scheduling based decision making in work environments where environment dynamics subject to numerous fluctuations frequently. Reinforcement Learning is an emerging field extensively research on environmental modelling under uncertainty. Optimization in dynamic scheduling can be effectively handled using Reinforcement learning. This paper presents a research that accompanies Multi-agent reinforcement learning (MARL) and extends Q-learning, which aims to provide a solution for the dynamic task scheduling in an uncertain environment. The proposed solution approaches as a model-based Deep Dyna-Q+ algorithm-based hybrid solution to solve the scheduling problem in an unexpected complex environment. The Dyna-Q+ based scheduler includes priority-indexing based task scheduler which allocates jobs to individual cooperative agents sequentially in adaptive mode. The proposed model-based solution is comparatively evaluated against benchmark methods; Q-Learning and Dyna-Q+ learning algorithms for dynamic task scheduling. The initial evaluation results verify that the approach would be an effective solution for scheduling optimization in a dynamic and complex environment.