The project is CLI to execute a GA for the scientific workflow scheduling problem.
To create the jar you have to issue.
mvn clean install
Then you can run:
java -jar target/sercheduler-0.0.1-SNAPSHOT.jar evaluate \
--workflowFile workflowFile \
--hostsFile hostsFile \
--seed seed \
--executions executions \
--fitness [simple,heft,rank,multi]
java -jar target/sercheduler-0.0.1-SNAPSHOT.jar evaluate --workflowFile src/test/resources/montage.json \
--hostsFile src/test/resources/hosts_test.json \
--seed 1 \
--executions 1000000 \
--fitness heft
We need a folder with the following structure
- benchmark-data
- workflows (insde all workflows)
- hosts
- mixed (inside all hosts)
- fast (inside all hosts)
java -jar target/sercheduler-0.0.1-SNAPSHOT.jar experiment \
-W workflowsPth \
-H hostsPath \
-T type \
--seed seed \
--executions executions
We can also do an analysis of multi fitness to get information about which method wins each round.
java -jar target/sercheduler-0.0.1-SNAPSHOT.jar analyze \
-W workflowsPth \
-H hostsPath \
-T type \
--seed seed \
--executions executions
For further reference, please consider the following sections: