DAISIM

The repository contains code for simulating a population of investors in the DAI Ecosystem.

• sim.py is a CLI to run a single MAKER DAO market simulation using a test config as input.

• simulation_util.py contains MAKER DAO market simulation code.

• single_user.py contains CVXPY optimization code for an optimal investor.

• test_runner.py is a test runner to run a MAKER DAO market simulations using multiple test configs.

• plot_gen.py generates multiple plots from the simulation output i.e. sim-summary.pickle.

• util.py contains all utility functions.

• input_generator.py is a CLI to generate test configs for a factorial experiment.

• Install cvxpy, pickle, numpy, scipy, matplotlib, tikzplotlib

Setup

Market Simulations

• A single market simulation takes in some inputs,

  • cdp_rate: CDP Rate for creating a MAKER CDP.
  • tx_fee: Transaction fee for buying/selling of ETH/DAI/cETH.
  • eth_price_feed: ETH Price over several days. For a n-day market simulation this is vector of size n containing ETH price for each day.
  • dai_price: Initial Price of DAI. Set to $1.
  • num_investors: Number of optimal investors participating in the market simulation.
  • assets_and_risk: Initial asset holdings and risk preference for all investors. This is a vector of size (num_investors,5) with each investors assets_and_risk a vector [USD, ETH, DAI, cETH, risk_param]. A lower numerical value for risk translates to high risk.
  • belief_factor: A constant indicating the strength of investors' belief that the price of DAI is 1.

• A sample config file is shown below for a MAKER DAO market simulation for a set of tx_fee and cdp_rate combinations. The config will be used to run 5 * 3 = 15 single market simulations with the given asset allocations and risk parameters for 4 investors.

2 7 0.01                                      // cdp_rate = [0.02, 0.03, 0.04, 0.05, 0.06]
5 8 0.01                                      // tx_fees = [0.05, 0.06, 0.07]
130                                           // ETH Price  (1-day market simulation)
4                                             // num_investors
252.3051 1.3159 278.8387 0.0 0.003            // Investor#0 Assets, Risk = 0.003
822.4563 2.0845 707.5078 0.0 0.003            // Investor#1 Assets, Risk = 0.003
399.4434 1.8082 459.3569 0.0 0.01             // Investor#2 Assets, Risk = 0.01
533.1002 2.3154 333.1751 0.0 0.01             // Investor#3 Assets, Risk = 0.01 
10                                            // belief_factor = 10

• Running MAKER DAO market simulations,
  • python3 sim.py --config path/to/config --logdir path/to/log/directory --days_per_config num_days_per_config : Running this generates a file sim-summary.pickle inside the log directory which is used to generate useful plots.
  • python3 plot_gen.py --data path/to/log/directory/sim-summary.pickle : Running this generates several useful plots for the simulation. All generated plots would show up in a plots directory under the log directory.
  • python3 test_runner.py --logdir /path/to/log/directory --configdir /path/to/config/directory : Running this performs market simulation with several test configs under a single directory i.e configdir.