Top five most promising algorithms in scheduling

This paper aims to be a short literature review, presenting the top five most promising algorithms for scheduling, as identified by us from the technical and scientific literature of the past years: Task Swap, Squeaky Wheel Optimization, Value-Biased Stochastic Search, Bee Colony Optimization And Temporal Difference (lambda), from reinforcement learning. We wanted to cover permutation-state methods, search-state methods, bias methods, swarm intelligence and machine learning. For accuracy, for each algorithm, we provide its description summarizing the original paper, and mention its strengths and weaknesses. Even if each algorithm may address particular issues, in order to prove their eligibility, but also to have an unified benchmark, we imagined an on-line oversubscribed scheduling scenario, named Simplified Automobile Repair Shop scheduling problem, and used data from a real automobile repair shop for testing.