A Case Study on Air Combat Decision Using Approximated Dynamic Programming

Algorithm 1

Utility function approximating based on sampled states.

ADP_Learn()

Input variables:

(1)   : sampled states set;

(2)  : the number of learning round.

(3)  : blue plane’s policy derived from Min-Max approach.

Output variables:

(1)   : utility function approximated.

Local variables:

(2)  : action vector of blue plane derived from Min-Max policy;

(3)  : action vector of red plane derived from current utility;

(4)  : one step improved utility function by Bellman iteration;

(5)  : features vector used to compute approximation coefficients;

(6)  : vector of approximation coefficients.

Code:

(1)   ;

(2)  FOR , DO:

(3)  ;

(4)  

(5)     ;

(6)  

(7)  

(8)  

(9)  END

(10) RETURN ;