how do I turn a piece of code into a parametric optimisation

I have a piece of code that I would like to turn into a parametric optimisation

clear
close all
%
% Matrix properties (8552 example)
Em = 380E3; % Young's modulus
num = 0.22; % Poisson's ratio
Vv = 0.01;
KIc = 3.63; % MPa*sqrt(m) fracture toughness
% Dv = 1E-3; % void diameter
%
% Crack density parameter (from Budiansky and O'Connel, IJSS 12:81-97, 1976)
cdp0 = 0;
cdp = cdp0;
a = 0.01;
%
%%
% MECHANISMS OF CYCLIC FATIGUE-CRACK PROPAGATION IN A FINE-GRAINED ALUMINA
% CERAMIC: THE ROLE OF CRACK CLOSURE C. J. Gilbert, R. O. Ritchie 2008
C = 1.3E-18; % Paris law coefficient
m = 23; % material exponent
%
%%
% some code goes here
for k = 1:N_cycles
    % some code goes here
end

I would like to change the variables Vv between 0.01, 0.38, and 0.47. I woul like to then be able to optimise a, C, and m so I can determine when these numbers give me the closest to arbitray values from literature.

Thanks

Alex

Answer 1

[Walter Roberson]

%
% Matrix properties (8552 example)
Em = 380E3; % Young's modulus
num = 0.22; % Poisson's ratio
Vv_vals = [0.01, 0.38, 0.47];
KIc = 3.63; % MPa*sqrt(m) fracture toughness
% Dv = 1E-3; % void diameter
%
% Crack density parameter (from Budiansky and O'Connel, IJSS 12:81-97, 1976)
cdp0 = 0;
cdp = cdp0;
a = 0.01;
%
%%
% MECHANISMS OF CYCLIC FATIGUE-CRACK PROPAGATION IN A FINE-GRAINED ALUMINA
% CERAMIC: THE ROLE OF CRACK CLOSURE C. J. Gilbert, R. O. Ritchie 2008
C = 1.3E-18; % Paris law coefficient
m = 23; % material exponent
%
%%
num_Vv = length(Vv_vals);
for Vv_idx = 1 : num_Vv
    Vv = Vv_vals(Vv_idx);

    % some code goes here
    for k = 1:N_cycles
        % some code goes here
    end
    results_per_Vv(Vv_idx) = appropriate per-Vv result
end

[A Poyser]

for the other variables will this nested loop also work?

[Walter Roberson]

Yes. Use a vector of actual values for each of them, and calculate length of the vector, and iterate 1 to length of the vector, extract the relevant value from the full list, and use the iteration index to index where to store results.

Note: it is advisable to pre-allocate the output variables; use the recorded lengths of the vectors to size the output variables.