Artificial intelligence for throughput bottleneck analysis – State-of-the-art and future directions

• Systematic literature review of relevant research efforts focusing on AI for throughput bottleneck analysis. Presented a classification structure four categories: identify, diagnose, predict and prescribe. For each category, outlined state-of-the-art re: input data, modelling approach outputs. Provided 8 future directions to advance the field Outlined 7 recommendations adoption academic solutions by manufacturing companies. Identifying, eventually eliminating bottlenecks, is key means increase productivity in production systems. In real world, however, bottlenecks challenge. This due landscape complex factory dynamics, with several hundred machines operating at any given time. Academic researchers have tried develop tools help identify eliminate bottlenecks. Historically, focused developing analytical discrete event simulation approaches However, rise industrial digitalisation artificial intelligence (AI), explored different ways which might be used based vast amounts digital shop floor data. By conducting systematic review, this paper aims present into use To make work more accessible practitioners, are classified (1) (2) (3) (4) was inspired real-world management practice. The categories, diagnose focus analysing historical whereas prescribe also provides topics practical may further push boundaries theoretical