gatraffic Is a Genetic Algorithm Traffic Agent for decentralized traffic management, the current implementation is used for testing purposes
There are two ways to testing the code, downloading directly the code from https://github.com/suarna/gatraffic.git or using a docker image downloaded from https://hub.docker.com/repository/docker/suarna/gatraffic
-Clone the repository:
git clone https://github.com/suarna/gatraffic.git
-Install sumo simulator from your repository if it is available for your OS, for example for Ubuntu Linux:
sudo apt-get install sumo sumo-tools
-Or download from http://sumo.dlr.de/docs/Downloads.php and follow the instructions for your OS:
-Set the environment variable $SUMO_HOME (if not yet) pointing to the sumo folder:
$SUMO_HOME=/usr/share/sumo
-The best way to satisfy the dependencies is creating a virtual environment into the project main folder, we need to satisfy some dependencies:
sudo apt-get install python3.8
sudo apt-get install python3-pip
sudo pip3 install virtualenv
-Finally we can create the environment:
python3.8 -m virtualenv venv
-Enable Python Virtual Environment using the following command:
source venv/bin/activate
-Install the required python packages using pip into the main :
pip3.8 install -r requirements.txt
-Execute the code:
python3.8 gatraffic.py
or
python3.8 test-gatraffic.py
-When finished experiments write deactivate in the prompt to leave the python virtual environment
-Install docker for your distribution from:
https://hub.docker.com/search?q=&type=edition&offering=community
-Open a terminal and follow the installation instructions.
-Download the image from the docker HUB repository:
sudo docker pull suarna/gatraffic:latest
-Run the image with the following command:
sudo docker run -it --name gatraffic suarna/gatraffic:latest bash
-If everything works fine, a prompt, something similar to this, should appear:
root@b0d1305bac4b:/app#
-Running the code executing the command:
python3.8 gatraffic.py
or
python3.8 test-gatraffic.py
-That's all for begin to experiment.
-Once the experiments have finished you can leave the container using
Ctrl+p followed by Ctrl+q combination, in this case the container remains active,
if you type exit the container will stop.
-To reconnect a running container
sudo docker exec -it gatraffic bash
-To connect as superuser and install extra applications you can do it using:
sudo docker exec -u 0 -it gatraffic bash
-If you want to restart the image type and attach the prompt:
sudo docker container start gataffic && sudo docker attach gataffic
-If you want to delete it once stopped ,simply write:
sudo docker container rm gatraffic
-If it is not stopped yet:
sudo docker container stop gatraffic && sudo docker rm gatraffic
-The parameters are changed editing the parameter files @param.txt and @param_test.txt, the nano utility is installed along with the docker image:
nano @param.txt or nano @param_test.txt
POPULATION_SIZE = 150
MU = 30
LAMBDA = 50
P_CROSSOVER = 0.8
P_MUTATION = 0.2
MAX_GENERATIONS = 2
HOF_SIZE = 6
CONFIG_FILE_ROUTE = "Nets/SimpleNet/sumo.sumocfg"
CYCLE = 120
MIN_PH_TIME = 15
MIN_OFFS_TIME = 0
MAX_OFFS_TIME = 100
INTERSECTION_ID = '101' `
CHROMOSOME_LENGTH = np_phases + 1
POPULATION_SIZE = 150
MU = 30
LAMBDA = 50
P_CROSSOVER = 0.8
P_MUTATION = 0.2
MAX_GENERATIONS = 3
HOF_SIZE = 6
RANDOM_SEED = rd.randint(0, 1000)
CONFIG_FILE_ROUTE = "Nets/SimpleNet/test-sumo.sumocfg"
N_EXPERIMENTS = 5
N_STEPS = n_steps
CYCLE = 120
MIN_PH_TIME = 15
MAX_PH_TIME = CYCLE-(MIN_PH_TIME*np_phases)
MIN_OFFS_TIME = 0
MAX_OFFS_TIME = 100
SINGLE = False
SEL_AL1 = False
SEL_AL2 = False
ADAPT = False
INTERSECTION_ID = '101'
BOUND = 0.02
-The parameters SINGLE, AL_1 and AL_2 when set to True are used to compute with other selection algorithms, SPEA2 and selNSGA3, selecting SINGLE we select the best individual.
-The bound parameter states how randomized the simulation is, 0,02 is a quite high value.
-Other parameter is related with verbosity of the output, we can change it in trafficinteract.py in the line 98 setting the value verbosity True or False
-We must also set some simulation parameters, we can edit the demandpedestrian.rou.xml
changing the flow probabilities and begin end simulation parameters, the same apply for
testdemandpedestrian.rou.xml.
**It is important that the simulation duration takes the same begin and en values for all the flows
The simulation results are stored in files in the folder data/, the output plots are stored into the folder
Net/SimpleNet/plots, in both cases the files are stored with names related with its content and date and time info.