List of standard modules and models

Here, we list all the modules and model types, that are shipped with the AgentLib core. The identifier specifies the string that should be used in type to use the module or model (see example at the Reference bottom of this page)

Communicators

link to simulator Reference simulator link to simulator Reference Simulator

Identifier	Extra dependencies	Description	Reference
local	-	Exchanges messages between agents executed within the same Python process. Usable with realtime and instant execution. Specify, which agents you subscribe to.	local
local_broadcast	-	Exchanges messages between agents executed within the same Python Process, realtime or instant execution. Messages are sent to all other agents with a local_broadcast communicator.	local_broadcast
multiprocessing_broadcast	-	Exchanges messages between all agents within a network that have this communicator, and are configured with the same ipv4 and port.	multiprocessing_broadcast
mqtt	mqtt (paho-mqtt)	Exchanges messages through MQTT. Requires the address of an MQTT broker, and you need to specify, which agents you subscribe to.	mqtt
clonemap	clonemap	Exchange messages through cloneMAP (with MQTT). Used, when MAS is executed via cloneMAP.	clonemap

Controllers

Identifier	Required dependencies	Description	Reference
pid	-	A simple PID controller.	pid
bangbang	-	A simple hysteresis controller.	bangbang

Miscellaneous

Identifier	Required dependencies	Description	Reference
simulator	-	A module to simulate models. Models can be FMU, state-space or custom defined, or come from plugins.	simulator
agentlogger	-	Logs the AgentVariables that are exchanged within an agent, effectively keeping a log of the state of an agent.	agent_logger
trysensor	-	Sends weather data of the Test Reference Year 2015.	try_sensor

Models

Identifier	Extra dependencies	Description	Reference
statespace	scipy	Linear state space model based on scipy.	fmu_model
fmu	fmu (fmpy)	Used to simulate FMUs.	scipy_model

Example

When using modules, they are identified by the "type" keyword in the configuration, which should include an identifier from this table, or a custom injection (combination of file and class name). For example, to run an agent with a simulator and an FMU model, the config should look like this:

{
  "id": "myAgent",
  "modules": [
    {
      "type": "simulator",
      "model": {
        "type": "fmu",
        "path": "path/toFMU.fmu"
      },
      
      "...": "remaining simulator config goes here"
    }
  }