---

---

1. Purpose

The Adaptive Score Function Estimator (ASFE) allows to estimate the sensitivities (derivatives) simultaneously with the expected performance of a Static Simulation System.

It is assumed that

The performance L of a system S in question is a deterministic function of a random state vector Y of certain dimension N:

L=L(Y),

and that the structure of this random vector is as follows:
 

[A:] all entries in Y are independent

[B:]

first n₁   entries in Y are distributed according to a given probability density function (pdf)  f₁

next n₂  entries in Y are distributed according to a given pdf  f₂

...

last n_k  entries in Y are distributed according to a given pdf  f_k  

 

The user is capable to provide a function Simulator which, given on input a state vector Y, returns on output the related performance L(Y) (see Section 5)

The user is capable to specify the distribution for each component of Y (see Section 4).

---

---