A Presentation of Several Simplex Optimization Techniques |
Contents
THE SIMPLEX METHOD OF NELDER AND MEAD | 5 |
REFLECTION IN A DIRECTION AS DETERMINED | 13 |
IV | 16 |
10 other sections not shown
Common terms and phrases
140 CONTINUE area of search Chebyquad COMMON/ONE/N CONTINUE RETURN END current simplex DETERMINES WHICH VERTEX DIMENSION PI(11 DIMENSION X(10 elements estimation evaluate f(x exponential weighting FORMAT 10X FORTRAN FUNC G EVAL GO TO 21 GO TO 64 GO TO 85 go to step gradient Hessian matrix HIGHEST FUNCTION VALUE initial simplex iteration linear search MATOUT 55 Mead's method method of Nelder method presented methods G MODBAD=1 MODOK=0 MODCHK.EQ.1 MODCHK=0 GO modification MUSOUT Nelder and Mead number of dimensions number of function objective function optimization techniques past samples PMIN PMIN=PINDEX presented in Chapter priori knowledge Proceed to step PSTORE 55 PSTORE INC1 quadratic approximations quadratic function rate of convergence reflection step RETURN END FUNCTION return to step Rosenbrock search algorithm search method set MODBAD equal SIMPLEX METHOD Singular SUBPROGRAM DETERMINES SUBPROGRAM PRODUCES Table TEST FUNCTION tion tive TTTT value of f(x vertices