Implementation of Estimation of Distribution Algorithms (EDAs).

EDA is a framework to optimize black-box discrete optimization problems.
The algorithm of EDA is as follows.

1. Initialize a population P whose size is λ.
2. Construct a population S which includes promising solutions in P.
3. Build an explicit probabilistic model M based on S.
4. Generate new λ_candidate solutions from M to construct a population O.
5. The solutions in P is replaced with those of O.
6. If termination conditions are met, then the algorithm is terminated, else go to (2).

EDAs requires:

• Python >= 3.6

Install EDAs from the sources:

git clone https://github.com/e5120/EDAs.git
cd EDAs
pip install -r requirements.txt
pip install -e .

(Optional) If you want to use main.py or eda/builder.py, you need to install BB-DOB project.

• PBIL
• UMDA
• CGA
• MIMIC
• ECGA
• AffEDA
• BOA

1. cd scripts
2. Rewrite a script file xxx.sh, if necessary. See output of python ../main.py -h for details of each parameter.
3. Execute a command bash xxx.sh.