This is a random search algorithm. For a description see Schumer, M.A. and K. Steiglitz, Adaptive Step Size Random Search, IEEE Transactions of Automatic Control, AC-13 (1968),270-276.
NQ = i2 = 4*NP
IER = Greater than 0 if optimum is achieved
IER = 3 Function value is accurate to relative ACC
IER = 1 Attempted stepsize is less than ACC (relatively)
IER = -1 Iteration limit exceeded
IER = -3 Input error
IER = -4 Not enough scratch storage
IER = -9 Function error in first FUNC call
IER = -15 No stopping criterion is enabled
SECTION 7.1 OPTIONS
Besides the general options the following are available.
DELTAZ = Initial stepsize (Default value is 4.0)
SDELTA = Minimum stepsize permitted. (Default value is 1.D-4). For effective optimization SDELTA should be .LE. 0.1*ACC
RHO = Stepsize shrinkage factor (default value is 0.5)
Note: In this algorithm the value of ITERL should be set at the maximum number of function evaluations permitted. The algorithm will compare ITERL with IVAL, not with ITERC. Iterations are still counted by ITERC but are of no interest to the general user.
Return to the Beginning