Comparative Study of Genetic Algorithm for Solving Teacher Placement Problem

Teacher placement is a combinatorial optimization problem in educational management that requires simultaneously satisfying teacher qualifications, school requirements, geographical constraints, and personal preferences.As the scale of educational systems grows, manual assignment becomes impractical, and the problem's NP-hard nature necessitates efficient computational approaches.This study applies a Genetic Algorithm (GA) framework to evaluate the effectiveness of four crossover operators, including Single-Point Crossover (SPX), Two-Point Crossover (TPX), Cycle Crossover (CX), and Ordered Crossover (OX) for solving the teacher placement problem.The GA uses permutation encoding, roulette wheel selection, and partial shuffle mutation, and operates on real-world data from the Magelang Regency Education Office, comprising 636 teachers and 106 schools.The objective is to minimize the total commuting distance between teachers' residences and assigned schools under varying mutation-to-crossover probability ratios (1:20 to 1:100).Experimental results show that OX consistently produces the best solutions, achieving the lowest average fitness value (10,301.63)across all configurations, followed closely by CX.In contrast, SPX and TPX demonstrate performance degradation at higher crossover probabilities, likely due to their inability to preserve valid permutations.Statistical analysis, including ANOVA and Kruskal-Wallis tests, confirms significant differences in performance, reinforcing the superiority of permutation-preserving crossovers.These results provide actionable guidance for designing intelligent teacher placement systems and selecting optimal GA operators for complex, real-world allocation problems.