stjordanis / tum_metaheuristics Goto Github PK

This project applies various metaheuristics to the airport gate assignment problem with time windows formulated by Lim et al in 2005 [1]. The study is conducted as part of the advanced seminar "Scheduling with Metaheuristics" at the chair of Operations Management of the TUM School of Management.

The following algorithms were implemented in Python 3.7.5 with numpy 1.17.4 and numba 0.46.0:

• Tabu Search
• Genetic Algorithm
• Memetic Algorithm
• Bee Colony Optimization

Their performance is compared to the commercial solver Gurobi.

Directory metaheuristics: Contains the implementation of Gurobi, Tabu Search, Bee Colony Optimization, Genetic and Memetic Algorithm. The instance generation and parameter setting is realized in problem.py. gantt.py enables the creation of a gantt chart for a given solution.

Directory experiments: Contains experimental_study.py which starts a new run of an experiment. The result data of our conducted experiment was saved to experimental_results.csv. Besides these files there are multiple files for visualizing our result data. Each plot has it own file.

[1] Lim, Andrew, Brian Rodrigues, and Yi Zhu. "Airport gate scheduling with time windows." Artificial Intelligence Review 24.1 (2005): 5-31.