zsliu98 / CSE6140-TSP

Gatech CSE6140/CX4140 Fall 2021 TSP Project

CSE6140-TSP

Gatech CSE6140/CX4140 Fall 2021 TSP Project

Code structure

• tsp_main.py: program entry, has four arguments
  • -inst: the path of the input dataset
  • -alg: the algorithm, from Branch-and-Bound / Approximation / Local Search 1 / Local Search 2
  • -time: cutoff time, in seconds
  • -seed: random seed
  • Usage:

    python3 tsp_main.py -inst <filename> -alg [BnB | Approx | LS1 | LS2] -time <cutoff_in_seconds> [-seed <random_seed>]

• eva_main.py: evaluation entry, has five arguments
  • -out: the evaluation method, from Comprehensive Table / QRTDs / SQDs / Box plots
  • -i: the input path
  • -o: the output path
  • -run: whether force it to rerun all algorithms
  • -alg: the algorithm, from Branch-and-Bound / Approximation / Local Search 1 / Local Search 2 / All algorithms
  • Usage:

    python3 eva_main.py -out [table | qrtd | sqd | box] -i <input_path> -o <output_path> -run [T | F] -alg [BnB | Approx | LS1 | LS2 | all]

  • evaluate/table.py: output comprehensive tables
  • evaluate/qrtd.py: output QRTD plots
  • evaluate/sqd.py: output SQD plots
  • evaluate/parameters.py: evaluation parameters
  • evaluate/check_run.py: run algorithms if there are no running results
• graph.py: load graph from datasets
• bnb.py: solve TSP with Branch-and-Bound
• approx.py: solve TSP with MST Approximation
• ls1.py: solve TSP with SA
• ls2.py: solve TSP with 2-opt
• output/: running results
• output/eva/: evaluate results