A stand alone nodejs app or module for creating robot trading simulations in a single market with configurable supply and demand.
By default uses my implementation of Gode and Sunder's Zero Intelligence Robots, but you can use other robot types or also write and register your own robots.
The ESDoc site for single-market-robot-simulator contains documentation prepared from source code of this module.
To run as a nodejs command-line program, clone this repository and run
npm install from the cloned
directory to install the dependencies:
git clone https://github.com/DrPaulBrewer/single-market-robot-simulator cd ./single-market-robot-simulator npm install
If, instead, you want to use it as a library in another program from npm, simply use
npm i as usual:
npm i single-robot-market-simulator -S
To use this as a library on the browser with jspm, you should set an override option on install forcing dependency
This is done in the robot-trading-webapp example app that uses this code as a dependency.
Configuration in the stand alone app occurs in the file
config.json which is read by
main() in stand-alone app mode.
When used as a software module, the configuration object is passed to
Format for config.json is given in configSchema.json as a JSON Schema.
When used as a stand alone app
node index.js or
npm run will run the simulation, reading the
config.json file and
outputting log files:
With the exception of
profits.csv these logs have header rows and are in comma-separated value format, compatible with
Excel and other spreadsheets and most analysis software that accepts a
.csv file as data input.
import * as SMRS from 'single-market-robot-simulator'; // ES6 const SMRS = require("single-market-robot-simulator"); // CJS
returns an object containing constructors for
Simulation and function
runSimulation. Simulation functionality
require different procedures for writing out files. Therefore, the log files are not immediately written out to .csv files
(as with the stand alone app) but are maintained in memory for use with browser-based plotting software. It is the
responsibility of other browser software (e.g.
single-market-robot-simulator-savezip) to write the logs to browser-side
Simulations can be run in either synchronous or asynchronous mode. Asynchronous mode is useful for running on the browser so that the event loop and user interface does not freeze while waiting for simulation results.
Example source code for a web-based simulator based on
single-market-robot-simulator may be found at
and the resulting simulator web app is at
will run the tests, if you have node 6 or later and mocha installed. You can also click on the build or coverage badges to view public test reports.
Copyright 2016 Paul Brewer, Economic and Financial Technology Consulting LLC