Abstract: With the annual rise in the number of creative startup enterprises in the Manufacturing Execution System (MES), there is
a rising need for research on the job-shop scheduling problem(JSSP). JSSP, a complex NP-hard issue, aims to optimize
allocating scarce resources to improve production efficiency. JSS is employed in several sectors, such as warehouse order
packaging and manufacturing on production lines. During practical circumstances, the operating context might become
Complicated owing to dynamic events such as the delivery of tasks at different times, delays in job completion, machine
malfunctions, or unexpected incidents. A lot of scheduling heuristics, including dispatching rules, have been used to
make good schedules by effectively prioritizing candidates, like manufacturing machines. Recently, Artificial Intelligence
(AI), specifically Deep Reinforcement Learning (DRL), has proven advantageous in autonomously acquiring scheduling
heuristics for JSS, mostly because of its versatility, effectiveness, and efficiency. Regrettably, there is an insufficient
number of surveys accessible to analyze and evaluate the benefits and drawbacks of these active initiatives. This work
seeks to address this deficiency by undertaking a thorough investigation of Bio-inspired heuristic algorithms and different
types of AI algorithms for JSS. This article has provided a concise overview of the utilized learning models, primary
objectives, datasets, and performance metrics. The outcomes are presented in tabular format, which enables a clear and
concise exhibition of the discoveries. In addition, this article analyzes certain issues and hurdles to recommend specific
areas that may be targeted for the future advancement of autonomous scheduling heuristic design.