Instead of combining the optimizer and solver information into a listing file that would be overly complex, a minimum of optimizer information is printed in the standard listing file. The standard listing file first displays the problem definitions, and then the data related to the optimization problem, as shown in the example that follows.
************************* * * * OPTIMIZATION * * * ************************* - Internal optimizer: Yes. - Sensitivities computed by Polyflow: Analytical. - Maximum number of solutions to reach optimum: 100 - Max iteration to find bracketed minimum: 3 - Max iteration before resetting search direction: 4 - Prec. for convergence of design variable: 0.1000000E-04 - Prec. for convergence of constraints: 0.1000000E-04 - Prec. for convergence of objective func.: 0.1000000E-04 - Initial penalty coefficient: 0.1000000E+01 - Penalty coefficient multiplier 0.2000000E+01 - Max penalty: 0.5000000E+06 DESIGN VARIABLES ---------------- * End_Parallel_Dis 0.0000000E+00 < s < 0.4990000E+01 Initial value = 0.0000000E+00 * J_F_Displ. 0.0000000E+00 < s < 0.1000000E+01 Initial value = 0.0000000E+00 OBJECTIVE FUNCTIONS ------------------- * Minimization of Objective Function #1 Multi-objective function priority : 0 F = X with X = FLOW_BALANCE CONSTRAINT FUNCTIONS -------------------- * G > 0 G = a*X + b with X = weighted average(PRESSURE) a = 0.1000000E+01 b = -0.1500000E+07
Information related to the current values that will be used for the next evaluation task are displayed in the next section of the standard listing file, as shown in the example that follows. These current values are listed before each evaluation of the solution or before the evaluation of the sensitivities.
Optimizer next task : Function evaluation Design Variables: next value - DV 1 : End_Parallel_Dis - DV 2 : J_F_Displ. Name | Cmp | Previous | Current | Target DV 1 | 1 | 0.9980000E+00 | 0.2612798E+01 | 0.2612798E+01 DV 2 | 1 | 0.1192911E+00 | 0.3123082E+00 | 0.3123082E+00
The following examples demonstrates what is displayed for an evaluation of the gradient:
Optimizer next task : Gradient evaluation Design Variables: next value - DV 1 : End_Parallel_Dis - DV 2 : J_F_Displ. Name | Cmp | Previous | Current | Target DV 1 | 1 | 0.3781609E+01 | 0.3781609E+01 | 0.3781609E+01 DV 2 | 1 | 0.4520164E+00 | 0.4520164E+00 | 0.4520164E+00 ANALYTICAL SENSITIVITIES COMPUTATION Solver : Sensitivities
After each evaluation of the solution, the values of the optimization functions (constraint and objective) are printed:
************************** * * * OPTIMIZATION * * * ************************** Optimizer current task : Function evaluation Design Variables: current value - DV 1 : End_Parallel_Dis - DV 2 : J_F_Displ. Name | Cmp | Previous | Current | Target DV 1 | 1 | 0.3346795E+01 | 0.3781609E+01 | 0.3781609E+01 DV 2 | 1 | 0.4000429E+00 | 0.4520164E+00 | 0.4520164E+00 Objective functions : - OF 1 : Objective Function #1 Name | Active | Value | Grad OF 1 | * | 0.3713507E+03 | Constraints : - CF 1 : Constraint on inlet pressure Name | Value | Grad CF 1 | -0.5551383E+05 |
After each evaluation of the sensitivities, the values of the optimization
functions (constraint and objective) and their sensitivities with respect to the
design variables are printed, as shown in the following example. Note that the
sensitivities are listed in the column labeled Grad
.
************************** * * * OPTIMIZATION * * * ************************** Optimizer current task : Gradient evaluation Design Variables: current value - DV 1 : End_Parallel_Dis - DV 2 : J_F_Displ. Name | Cmp | Previous | Current | Target DV 1 | 1 | 0.3781609E+01 | 0.3781609E+01 | 0.3781609E+01 DV 2 | 1 | 0.4520164E+00 | 0.4520164E+00 | 0.4520164E+00 Objective functions : - OF 1 : Objective Function #1 Name | Active | Value | Grad OF 1 | * | 0.3713507E+03 | DV 1 | | | -0.8564837E+01 DV 2 | | | -0.3527887E+02 Constraints : - CF 1 : Constraint on inlet pressure Name | Value | Grad CF 1 | -0.5551383E+05 | DV 1 | | -0.4006829E+06 DV 2 | | 0.4279410E+06
Finally, the total number of evaluations of the solution (including those that succeeded and failed) are printed at the end of the file, along with the number of evaluations of the sensitivities (that is, the gradient evaluations).
RESULT OF THE OPTIMIZER: ======================== _____________________ Number of function evaluation : 45 Number of gradient evaluation : 13 Optimization succeeded. GLOBAL CONVERGENCE: End of Optimization