Skip to content

config_docs.md

Latest commit

 

History

History
123 lines (93 loc) · 2.88 KB

config_docs.md

File metadata and controls

123 lines (93 loc) · 2.88 KB

Config File Documentation

This document outlines the structure and valid values for the config.yaml file. Each section below corresponds to a configuration parameter.

option_type

  • Type: string
  • Options:
    • "AMERICAN" — Option can be exercised at any time before expiration.
    • "EUROPEAN" — Option can be exercised only at expiration.
    • "BERMUDAN" — Option can be exercised at predefined points (see exercise_frequency).
  • Default: "BERMUDAN"

option_side

  • Type: string
  • Options:
    • "PUT" — Grants the right to sell an asset at the strike price.
    • "CALL" — Grants the right to buy an asset at the strike price.
  • Default: "PUT"

dimensions

  • Type: int
  • Specifies the number of underlying assets to price. Put 1 for vanilla options
  • Default: 5

risk_free_interest

  • Type: float
  • Symbol: r
  • In a risk-neutral framework, the risk-free rate also serves as the expected return (μ).
  • Default: 0.05

time_to_exp

  • Type: float
  • Symbol: T
  • Time to expiration, measured in years.
  • Default: 0.25

exercise_frequency

  • Type: string
  • Only applicable if option_type = "BERMUDAN".
  • Options:
    • "MONTHLY"
    • "QUARTERLY"
    • "SEMI_MONTHLY"
    • "CUSTOM" — Requires the custom_exercise_points parameter.
  • Default: "MONTHLY"

exercise_points

  • Type: Array[int]
  • Only applicable if option_type = "BERMUDAN" and exercise_frequency = "CUSTOM".
  • Specifies custom exercise times in days.
  • Example: [1, 50, 100, 300] — each value represents a day on which early exercise is allowed.

init_stock_prices

  • Type: Array[float]
  • Symbol: [S₀₁, S₀₂, ..., S₀_d]
  • Initial stock prices for each asset.
  • Array length must equal dimensions.
  • Default: [80, 82, 78, 81, 83]

strike_prices

  • Type: Array[float]
  • Symbol: [K₁, K₂, ..., K_d]
  • Strike price for each asset.
  • Array length must equal dimensions.
  • Default: [110, 110, 110, 110, 110]

volatilities

  • Type: Array[float]
  • Symbol: [σ₁, σ₂, ..., σ_d]
  • Volatility of each asset.
  • Array length must equal dimensions.
  • Default: [0.1, 0.12, 0.09, 0.11, 0.1]

num_of_paths

  • Type: int
  • Symbol: M
  • Number of Monte Carlo simulation paths.
  • Default: 10000

num_of_steps

  • Type: int
  • Symbol: N
  • Number of discrete time steps used in each path.
  • Default: 250

poly_degree

  • Type: int
  • Degree of the polynomial used in the regression step of the Longstaff-Schwartz method (LSM).
  • Default: 3

epochs

  • Type: int
  • Number of training epochs for the feedforward neural network (FNN) when used.
  • Default: 300