forward.impulse.policy: Simulate a payoff of an impulse strategy along a set of...

In mludkov/mlOSP: Machine Learning and Regression Monte Carlo Algorithms for Optimal Stopping

Simulate a payoff of an impulse strategy along a set of forward paths

Description

Simulate a payoff of an impulse strategy along a set of forward paths

Usage

forward.impulse.policy(x, M, fit, model, mpc = FALSE)

Arguments

x	is a matrix of starting values if input x is a list, then use the grids specified by x
M	number of time steps to forward simulate
fit	a list of M fitted emulators that determine the functional approximators of V(k,x). Supports km, spline, and hetGP objects (anything supported by ospPredict)
model	a list containing all model parameters. In particular need model$impulse.func for computing the intervention operator (optimal impulse to consider), model$sim.func for simulating each step with time step model$dt.

Details

Should be used in conjunction with the osp.impulse.control function that builds the emulators and calls forward.impulse.policy internally.

Value

a list containing:

• payoff (vector) is the resulting payoff NPV from t=0

• tau (vector) number of times impulses were applied on each path

• impulses (matrix) impulse amounts matching tau

• paths ((d+2)-tensor) forward trajectories of the controlled state process

• bnd (vector) impulse target levels for the case of linear impulse costs

mludkov/mlOSP documentation built on April 29, 2023, 7:56 p.m.