A Systematic Literature Review of Metaheuristic Algorithms for Course, Exam and School Timetabling

Abstract

Timetabling (Course, exam and school timetabling) remains a central combinatorial optimization problem with broad practical importance. In the last decade metaheuristics have continued to dominate applied solutions, and recent years show renewed interest in hyper-heuristics and metaheuristics for robustness and scalability. This systematic literature review (SLR) restricts its scope to metaheuristic methods published in the last three years and limited to course, exam and school timetabling applications. We screened multiple bibliographic sources and selected fourteen qualifying studies (peer-reviewed articles, conference papers and high-quality preprints). The SLR synthesizes algorithmic families (genetic algorithms, tabu search, simulated annealing, particle swarm optimization, memetic algorithms, hyper-heuristics, metaheuristics and hybrids), highlights prominent design patterns (Graph-colouring initializers, multi-neighbourhood local search, high-level/low-level hyper-heuristics, surrogate-assisted metaheuristics), and discusses benchmarking practice and reproducibility. Results indicate (1) continued prevalence of genetic algorithms and local-search hybrids, (2) rising use of hyper-heuristic and surrogate-accelerated metaheuristic designs, and (3) heterogeneous benchmarking and limited code sharing that hinder direct comparisons. We conclude with recommended best practices for benchmarking and reproducibility and identify promising research directions, including dynamic/daily timetabling, adaptive operator selection, and rigorous statistical reporting.