Pareto frontier plot error

3 次查看（过去 30 天）
Briana Canet
Briana Canet 2023-9-18
评论： Torsten 2023-9-18
Ran in:
I am trying to plot the Pareto frontiers for...
Problem A
Minimize f1(x1,x2)=x2*sin(x1)
Minimize f2(x1,x2)=x1+x2
and
Problem B
Minimize f1(x1,x2)=x1^4+(4*x1*x2)
Minimize f2(x1,x2)=x1+(2*x2^2)
I am not sure when a dot/period is needed in f1 and f2 in Lines 2 and 3 for Problem A and B. I am only getting one point on (0,0) for each problem now.
Problem A:
% Functions
f1 = @(x1, x2) x2.*sin(x1);
f2 = @(x1, x2) x1+x2;
% Grid of values
[x1, x2] = meshgrid(0:0.01:1, 0:0.01:1);
% Calculate the objective functions
F1 = f1(x1, x2);
F2 = f2(x1, x2);
% Weighted average method
weights = 0:0.01:1;
paretoFronts = zeros(length(weights), 2);
for i = 1:length(weights)
l = weights(i);
% Weighted sum
F = l*F1 + (1-l)*F2;
[minF, idx] = min(F(:));
paretoFronts(i, :) = [x1(idx), x2(idx)];
end
figure;
plot(paretoFronts(:,1), paretoFronts(:,2), 'o-');
xlabel('x1');
ylabel('x2');
title('Problem 1ii');
grid on;
Problem B:
% Functions
f1 = @(x1, x2) x1.^4+(4*x1.*x2);
f2 = @(x1, x2) x1+(2*x2.^2);
% Grid of values
[x1, x2] = meshgrid(0:0.01:1, 0:0.01:1);
% Calculate the objective functions
F1 = f1(x1, x2);
F2 = f2(x1, x2);
% Weighted average method
weights = -1:0.01:1;
paretoFronts = zeros(length(weights), 2);
for i = 1:length(weights)
l = weights(i);
% Weighted sum
F = l*F1 + (1-l)*F2;
[minF, idx] = min(F(:));
paretoFronts(i, :) = [x1(idx), x2(idx)];
end
figure;
plot(paretoFronts(:,1), paretoFronts(:,2), 'o-');
xlabel('x1');
ylabel('x2');
title('Problem 1ii');
grid on;

请先登录，再回答此问题。

回答（1 个）

Torsten
Torsten 2023-9-18
移动：Torsten 2023-9-18
In both examples, the two objective functions are not "competitive" since both have (0/0) as optimum combination for x1 and x2 in the range 0 <= x1 <= 1 and 0 <= x2 <= 1. These are bad examples for pareto optimum problems.
  2 个评论
Briana Canet
Briana Canet 2023-9-18
Although it's not the best example, are the functions f1 anf f2 written correctly with appropriate placement of "."?
Torsten
Torsten 2023-9-18
Yes. As long as x1 and x2 have the same size (which is guaranteed since you generate them using "meshgrid"), the "."'s are appropriately placed.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Multiobjective Optimization 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by