Dealing with Error "Not enough Input Arguments" when using NONLCON for fmincon

Hello,

I'm trying to solve a basic non linear programming problem. It involves several non linear inequality constraints. I keep getting the error "Not enough input arguments" in the function fun3 whenever I try to run my full script. I'm new to matlab, so I know it is probably something very simple but I am stumped. Also, I probably misused the terminology in this post and I apologize.

Here is my code:

      %%main function
      clc
      clear
      tic

      Budget=(1000:1000:5000);
      SI=(100:100:1000);
      ST=(10:10:100);
      SA=(10:10:100);
      Elapsed_Time=zeros(5,10,10,10);

      for b=1:5
          TBudget=Budget(b);
          for i=1:10
              si=SI(i);
              for j=1:10
                  st=ST(j);
                  for k=1:10
                      sa=SA(k);

                      stime=zeros(st,sa);
                      x=zeros(st,sa);
                      y=zeros(2*st,2*sa);
                      target_recovery=0.3;

                      beta_l1=0.4;
                      beta_l2=0;
                      beta_l3=0.5;
                      alpha_r1=0.05;
                      alpha_r2=0.5;
                      alpha_r3=0.05;
                      alpha_r4=0.3;
                      delta_scost=0.3;
                      beta_cost=5;
                      lower_scost=6;
                      delta_ptime=5;
                      u=0.7;
                      upper_ptime=8;

                      lower_bound=0;
                      upper_bound=1;

                      p=0;
                      q=0.003;

                      for l=1:st
                          for m=1:sa
                              lb=zeros(2*st,sa);
                              ub=ones(2*st,sa);
                              for n=1:si
                                  [cost(l,m),recovery_function(l,m),loss_function(l,m),ptime(l,m),ctime(l,m),quality_response_function(l,n,m)]=fun2(stime(l,m),x(l,m),delta_scost,beta_cost,lower_scost,alpha_r1,alpha_r2,alpha_r3,alpha_r4,beta_l1,beta_l2,beta_l3,upper_ptime,delta_ptime,u,p,q);
                                  y(l,m)=stime(l,m);
                                  y(st+l,m)=x(l,m);
                              end
                          end
                          rtim=min(ctime(l,:));
                      end

                      dec1(:,1)=sum((recovery_function-loss_function),2);
                                  dec1(:,2)=sum(cost,2);
                      dec1(:,3:2+si)=sum(quality_response_function,3);

                          %optimal solution
                            y=fmincon(@fun1,rand(st,sa),[],[],[],[],lb,ub, fun3);

                  end
              end
          end
      end
      toc

      %%fun1
      function f=fun1(rtim)
      obj=sum(rtim,2);
      f=obj(1);

      %%fun2
      function [cost,recovery_function,loss_function,ptime,ctime,quality_response_function]=fun2(stime,x,delta_scost,beta_cost,lower_scost,alpha_r1,alpha_r2,alpha_r3,alpha_r4,beta_l1,beta_l2,beta_l3,upper_ptime,delta_ptime,u,p,q)

      cost= delta_scost*exp(-beta_cost*x)+lower_scost + x;
      recovery_function=alpha_r1*(1-alpha_r2*stime)+alpha_r3*(1-exp(-alpha_r4*x));
      loss_function=beta_l1*(1+beta_l2*stime)*exp(-beta_l3*x);
      ptime=upper_ptime-delta_ptime*(1-exp(-u*x));
      ctime=stime+ptime;

      quality_response_function= p*stime + q*x;

      %%fun3
      function [c,ceq]=fun3(dec1,TBudget,st,si)

      c=[-dec1(:,1)+0.3*ones(st,1);
      dec1(:,2)-TBudget*ones(st,1);
      -dec1(:,3:(si+2));
      dec1(:,3:(si+2))-ones(st,si)];

      ceq=[];

Thank you!

Answer 1

[Alan Weiss]

In addition to what Matt J said, your nonlinear constraint function needs to be a function of one input variable. Either take the function as a function of dec1 alone:

 @(dec1)fun3(dec1,TBudget,st,si)

or, if you have all four variables in the problem, bundle them into a new variable x as follows:

 function [c,ceq]=fun3(x)
 dec1 = x(1);
 TBudget = x(2);
 st = x(3);
 si = x(4);
 ...

Alan Weiss

MATLAB mathematical toolbox documentation

[Ruby]

It helps a lot.Thank you !

Answer 2

[Matt J]

You need to pass '@fun3' instead of just 'fun3':

    y=fmincon(@fun1,rand(st,sa),[],[],[],[],lb,ub, @fun3);

[Ruby]

I 've noticed that but it's not the most important reason the coding doesn't work. I've already finished it and thank you all the same!

[Matt J]

It is the answer to your question, however. The error message that you posted "Not enough input arguments..." and which was the basis for your posted question was due to that.