SEAGE is a framework for metaheuristic collaboration. See more at https://www.seage.org

Simply run the build script:

./scripts/build.sh

Then run the launcher which will provide you some help:

./scripts/run.sh

To see what is implemented run:

./scripts/run.sh info

To see what is implemented in more details for a specified problem run:

./scripts/run.sh details -p SAT

Traveling Salesman Problem (TSP) solved by the Genetic Algorithm:

./scripts/run.sh experiment-approach -a GeneticAlgorithm -i TSP:berlin52 -t 10 -n 10

Satisfiability Problem (SAT) solved by the Tabu Search

./scripts/run.sh experiment-approach -a TabuSearch -i SAT:uf20-01 -t 10 -n 10

When running with no database uri specified, the results are stored in the local H2 database. You can inspect the results as follows:

java -jar seage-launcher/build/install/seage-launcher/lib/h2-*.jar

Provide the following information:

• JDBC URL:jdbc:h2:<path-to-seage>/output/seage.local.h2
• username: sa
• password: empty

DB_USER=<username> \
DB_PASSWORD=<password> \
DB_URL="jdbc:postgresql://<server>/seage" \
./scripts/run.sh experiment-single-random -a TabuSearch -i SAT:uf20-01 -t 10 -n 10

You can run SEAGE also from source against the database running in docker. There are three containers:

• seage (you don't need to run it)
• database (postgresql, localhost:25432)
• pgadmin (http://localhost:28080)

docker compose up database pgadmin

DB_URL="jdbc:postgresql://localhost:25432/seage" ./scripts/run.sh experiment-single-random -a TabuSearch -i SAT:uf20-01 -t 10 -n 10

The results can be inspected using pgAdmin at the address http://localhost:28080

• username: [email protected]
• password: seage
• database password: seage

DB_URL=jdbc:postgresql://snoopy.lan:15432/seage ./seage report