writingpeter/coil_gecco23
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
Repository files navigation
README ------ Code for the paper entitled "Using a Variational Autoencoder to Learn Valid Search Spaces of Safely Monitored Autonomous Robots for Last-Mile Delivery" Please cite: Peter J. Bentley, Soo Ling Lim, Paolo Arcaini and Fuyuki Ishikawa (2023). Using a Variational Autoencoder to Learn Valid Search Spaces of Safely Monitored Autonomous Robots for Last-Mile Delivery. In Genetic and Evolutionary Computation Conference (GECCO’23). ACM, Lisbon, Portugal. Top-level directory . ├── Experiments # Experiments └── README.txt # README file Required packages: deap==1.3.1 pytorch==1.9.0 numpy==1.18.5 matplotlib==3.4.3 ---- Directory: Experiments ---- . ├── ... ├── Experiments # COIL code, standard GA code and experiment settings │ ├── config_t30_r10_d10000_l60.py # Settings for E1.1: COIL vs GA (also acts as a baseline for all other experiments) │ ├── config_t30_r15_d10000_l60.py # Settings for E1.2: Number of simultaneously running robots RT │ ├── config_t30_r20_d10000_l60.py # Settings for E1.2: Number of simultaneously running robots RT │ ├── config_t20_r10_d10000.py # Settings for E1.3: Total number of robots Rb │ ├── config_t25_r10_d10000.py # Settings for E1.3: Total number of robots Rb │ ├── max_req_durations # COIL code, standard GA code and experiment settings for E1.4: Changing request durations dr │ ├── config_t30_r10_d2500_l60.py # Settings for E2.1: Reducing dataset size DS │ ├── config_t30_r10_d5000_l60.py # Settings for E2.1: Reducing dataset size DS │ ├── config_t30_r10_d7500_l60.py # Settings for E2.1: Reducing dataset size DS │ ├── config_t30_r10_d10000_l10.py # Settings for E2.2: Modifying number of latent variables maxlv │ ├── config_t30_r10_d10000_l20.py # Settings for E2.2: Modifying number of latent variables maxlv │ ├── config_t30_r10_d10000_l30.py # Settings for E2.2: Modifying number of latent variables maxlv │ ├── config_t30_r10_d10000_l40.py # Settings for E2.2: Modifying number of latent variables maxlv │ ├── config_t30_r10_d10000_l50.py # Settings for E2.2: Modifying number of latent variables maxlv │ ├── generate_data.py # COIL Step 1: generate data │ ├── learn_representation.py # COIL Step 2: learns representation │ ├── optimise.py # COIL Step 3: optimise │ ├── ga.py # Standard GA │ ├── scheduler # Scheduler code used by COIL and GA │ ├── analyse.py # Compares results from GA and COIL │ ├── data # Folder containing data generated by generate_data.py │ ├── vae # Folder containing VAEs generated by learn_representation.py │ └── results # Folder containing results generated by optimse.py and ga.py └── ... * To run Experiment 1.1 (all the other experiments are run the same way): Running COIL: >> python generate_data.py -c config_t30_r10_d10000_l60 -t >> python learn_representation.py -c config_t30_r10_d10000 -t >> python optimise.py -c config_t30_r10_d10000_l60 -t -r 100 Running standard GA: >> python ga.py -c config_t30_r10_d10000_l60 -t -r 100 Comparing COIL with GA: >> python analyse.py -c config_t30_r10_d10000_l60