collaborative-learning

How to build

0. Assume $PROJECT_PATH contains full path to collaborative-learning folder and $OMNETPP_PATH to the Omnet++ one;

1. Create the environment for the python agent by invoking the following command:

conda env create -f $PROJECT_PATH/agent/environment.yml -p $PROJECT_PATH/agent/.conda/collaborative-learning-dev

2. Save in the environment variable AGENT_PATH the absolute path to the agent dir:

export AGENT_PATH=$PROJECT_PATH/agent

• checkout the cmake submodule

git submodule update --init --recursive

• set Omnet++ environment:

source $OMNETPP_PATH/setenv

3. Configure the cmake project and generate the Makefile:

/usr/bin/cmake --no-warn-unused-cli -DCMAKE_BUILD_TYPE:STRING=Debug -DCMAKE_EXPORT_COMPILE_COMMANDS:BOOL=TRUE -DCMAKE_C_COMPILER:FILEPATH=/usr/bin/gcc -DCMAKE_CXX_COMPILER:FILEPATH=/usr/bin/g++ -S$PROJECT_PATH -B$PROJECT_PATH/bin -G "Unix Makefiles"

4. Launch the simulation by building the Makefile:

/usr/bin/cmake --build $PROJECT_PATH/bin --config Debug --target run_collaborative-learning-sim -j 14

Analyzing results

Statistics are recorderd in the simulations/results folder.

To analyze statistics, open the simulations/results/General.anf in the Omnet++ IDE and use its Analysys Tool.

Agent configuration

The agent configuration is done through a JSON file named agent_conf.json. This file contains the basic configuration, which can be overridden in various ways:

• by modifying the file directly,
• by creating another file and specifying its path in the agent bean (in this case, all fields must be filled),
• by passing a new JSON string directly into the agent bean that specifies all or some of the parameters, with any missing parameters being taken from the base configuration file.

Documentation on how to configure the agent can be found at: Agent Configuration.