optim.prob​lemdef.Opt​imizationP​roblem/sol​ve SOLVE requires a non-empty initial point structure to solve a nonlinear problem.

Hi, I am currently creating an optimization problem to find the minimum value of an equation with 5 constraints. My code is:

clear;
clc;
prob = optimproblem("Description","BestWorstMethod");
A_B = [2 1 3 3 8];
A_W = [4;8;4;3;1];
n=length(A_B);
W = optimvar("W",n,"LowerBound",0);
e = optimvar("e",1,"LowerBound",0);
w_b = W(2);
w_w = W(5);
prob.Objective = e;
[cons1,cons2] = fcn2optimexpr(@consBO,A_B,W,w_b);
[cons3,cons4] = fcn2optimexpr(@consOW,A_W,W,w_w);
prob.Constraints.cons1 = cons1 <= e;
prob.Constraints.cons2 = cons2 <= e;
prob.Constraints.cons3 = cons3 <= e;
prob.Constraints.cons4 = cons4 <= e;
prob.Constraints.cons5 = sum(W)==1;
[sol,optimval] = solve(prob);

Solving problem using fmincon.

Error using optim.problemdef.OptimizationProblem/solve
SOLVE requires a non-empty initial point structure to solve a nonlinear problem.

optimW = sol.W

function [cons1,cons2] = consBO(A_B,W,w_b)
n=length(A_B);
cons1 = zeros(1,n);
cons2 = zeros(1,n);
for j = 1:n
    cons1(j) = w_b - A_B(j) * W(j);
    cons2(j) = A_B(j) * W(j) - w_b;
end
end
function [cons3,cons4] = consOW(A_W,W,w_w)
n=length(A_W);
cons3 = zeros(1,n);
cons4 = zeros(1,n);
for j = 1:n
    cons3(j) = W(j) - A_W(j) * w_w;
    cons4(j) = A_W(j)* w_w - W(j);
end
end

The function consBO and consOW define the for loop structure on constraints.The constraint inequality is shown below.

I keep getting the same errors:

Error using optim.problemdef.OptimizationProblem/solve

SOLVE requires a non-empty initial point structure to solve a nonlinear problem.

Please let me know if you have any advice. Thanks!

Answer 1

[Torsten]

My advice is: Do what the error message tells you to do and supply an initial point structure to make MATLAB happy.

initial.e = 0;
initial.W = ones(n,1);
[sol,optimval] = solve(prob,initial);