sanity checks implemented. For example, the following line of code is considered a
“success,” no matter if the content is indeed an instance of a discounting class or not.
Therefore, one has to be rather careful when compiling and passing market_environment
objects to any simulation class:
self.discount_curve = mar_env.get_curve(‘discount_curve’)
Table 16-1 shows all components that a market_environment object must contain for the
generic and therefore for all other simulation classes.
Table 16-1. Elements of market environment for all simulation classes
Element Type Mandatory Description
initial_value
Constant
Yes
Initial value of process at pricing_date
volatility
Constant
Yes
Volatility coefficient of process
final_date
Constant
Yes
Simulation horizon
currency
Constant
Yes
Currency of the financial entity
frequency
Constant
Yes
Date frequency, as pandas freq parameter
paths
Constant
Yes
Number of paths to be simulated
discount_curve
Curve
Yes
Instance of constant_short_rate
time_grid
List
No
Time grid of relevant dates (in portfolio context)
random_numbers
List
No
Random number array (for correlated objects)
cholesky_matrix
List
No
Cholesky matrix (for correlated objects)
rn_set
List
No
dict object with pointer to relevant random number set
Everything that has to do with the correlation of model simulation objects is explained in
subsequent chapters. In this chapter, we focus on the simulation of single, uncorrelated
processes. Similarly, the option to pass a time_grid is only relevant in a portfolio context,
something also explained later.
