Create a flatland environment for agent navigation.

Options

• :id - the ID for the agent (default: random UUID)
• :hidden_neurons - number of hidden neurons (default: 5)
• :world_size - size of the flatland world (default: {10, 10})

Examples

iex> Bardo.API.create_flatland()
{:ok, agent_id}

Create and configure an agent for algorithmic trading.

Options

• :id - the ID for the agent (default: random UUID)
• :instrument - trading instrument (default: "EUR_USD")
• :timeframe - trading timeframe (default: "M15")
• :hidden_neurons - number of hidden neurons (default: 10)

Examples

iex> Bardo.API.create_trading_agent(instrument: "BTC_USD")
{:ok, agent_id}

Create and setup a simple XOR neural network.

Options

• :id - the ID for the agent (default: random UUID)
• :hidden_neurons - number of hidden neurons (default: 2)
• :bias - whether to use bias neurons (default: true)
• :activation - activation function to use (default: :tanh)

Examples

iex> Bardo.API.create_xor()
{:ok, agent_id}

Deploy an agent for live trading.

Options

• :broker - the broker to use (default: :oanda)
• :config - broker-specific configuration

Examples

iex> config = %{api_key: "your_api_key", account_id: "your_account_id"}
iex> Bardo.API.deploy_trading_agent(agent_id, broker: :oanda, config: config)
{:ok, deployment_id}

Get performance metrics for an agent.

Examples

iex> Bardo.API.get_agent_metrics(agent_id)
{:ok, %{
  sharpe_ratio: 1.2,
  max_drawdown: 0.15,
  total_return: 0.25,
  win_rate: 0.55
}}

Initialize a new Bardo environment with sensible defaults.

Examples

iex> Bardo.API.init()
:ok

Load a pre-trained trading agent from a file.

Examples

iex> Bardo.API.load_trading_agent("/path/to/agent.json")
{:ok, agent_id}

Run a standalone training session for a trading agent.

Options

• :instrument - trading instrument (default: "EUR_USD")
• :timeframe - trading timeframe (default: "M15")
• :start_date - historical data start date (default: one year ago)
• :end_date - historical data end date (default: current date)
• :generations - number of generations to train (default: 100)
• :population_size - size of the population (default: 50)

Examples

iex> Bardo.API.run_standalone_training(instrument: "BTC_USD")
{:ok, %{agent_id: agent_id, fitness: 0.85}}

Save a trained agent to a file.

Examples

iex> Bardo.API.save_agent(agent_id, "/path/to/save/agent.json")
:ok

Stop a deployed trading agent.

Examples

iex> Bardo.API.stop_trading_agent(deployment_id)
:ok

Test the agent on the XOR problem.

Examples

iex> {:ok, agent_id} = Bardo.API.create_xor()
iex> Bardo.API.train_xor(agent_id)
iex> Bardo.API.test_xor(agent_id)
{:ok, %{
  inputs: [[0, 0], [0, 1], [1, 0], [1, 1]],
  outputs: [[0.02], [0.98], [0.97], [0.03]],
  expected: [[0], [1], [1], [0]]
}}

Train a flatland agent to navigate and avoid obstacles.

Options

• :generations - number of generations to train (default: 500)
• :population_size - size of the population (default: 50)
• :fitness_target - target fitness to reach (default: 800)
• :simulation_steps - steps per evaluation (default: 100)

Examples

iex> {:ok, agent_id} = Bardo.API.create_flatland()
iex> Bardo.API.train_flatland(agent_id)
{:ok, %{best_fitness: 825, generations: 320}}

Train an agent on the XOR problem.

Options

• :generations - number of generations to train (default: 100)
• :population_size - size of the population (default: 20)
• :fitness_target - target fitness to reach (default: 3.9)

Examples

iex> {:ok, agent_id} = Bardo.API.create_xor()
iex> Bardo.API.train_xor(agent_id)
{:ok, %{best_fitness: 3.95, generations: 42}}