GSACO

This repository demonstrates the application of ant colony foraging behavior to optimize the scheduling process within large-scale semiconductor manufacturing environments.

Datasets

For validation purpose, the datasets are available in "benchmark_instances" directory.

An ant colony algorithm implemented in Python using NetworkX for benchmark instances is available in "benchmark_instances/aco_networkX" directory. The constraint programming (CP) model used for benchmark instances is available in "benchmark_instances/cp" directory. It is derived from an open source and can be found here: CP/or-tools for FJSP

The large-scale semiconductor instances from SMT2020 is available in "HVLM" and "LVHM" directory.

The example instance used in paper for demonstration of algorithm working is available in "LVHM-ex" directory.

The CP model for SMT2020 instances is available in "cp_SMT2020" directory.

An ant colony algorithm for SMT2020 instances implemented in Python using PyTorch is available in "aco_cpu" directory

Installation

Install Python interpreter (suggested: 3.9)

python3 -m pip install -r requirements.txt

Usage

GSACO algorithm:

Reference directory: aco_cpu

Execute:

For windows:

main.py

For linux:

python main.py

For experiments, parameters of model can be adjusted within "get_input.py" file or command-line arguments.

CP model:

Reference directory: cp_SMT2020

Execute:

For windows:

cp_model.py

For linux:

python cp_model.py

The parameters can be adjusted within "cp_model.py" file (n, dataset).