Input arguments must be convertible to floating-point numbers

T=1;
N=10;
K=1000;
bar_W=20;
d=zeros(T,N,K);

mu=zeros(N,1);
for n=1:N
    mu(n)=30+(60-30)*rand();
end

sigma=zeros(N,1);
for n=1:N
    sigma(n)=mu(n)*0.3*rand();
end

for i=1:N
    d(T,i,:)=normrnd(mu(i),sigma(i), K,1);
end

[xij,seq]=X_IJ(T,N,bar_W,mu,sigma);

[a_talt,ref_bar_fix]=bar_fix(T,N,K,d,bar_W,xij);
 for i=1:N-1
     a_talt_total(i)=a_talt_total(i)+a_talt(i);
 end
 
 [a_opt,ref_heuristic]=heuristic(T,N,K,d,bar_W,xij);
 for i=1:N-1
     a_opt_total(i)=a_opt_total(i)+a_opt(i);
 end
 
 a_halt=zeros(N-1,1);    
 
 for i=1:N-1
     syms a_halt_copy
     tk=0;
     for k=1:K
          if i==1
              tk=tk+max(-(d(T,seq(i),k)-a_halt_copy),0);
          else
              tk=tk+max(-(bar_W+d(T,seq(i),k)-a_halt_copy),0);
          end
     end
     a_halt_copy=solve(a_halt_copy-mu(seq(i))-sum(sum(sum(tk)))/K==0,a_halt_copy);
     a_halt(i)=a_halt_copy;
 end

Hi everyone, im getting this error:

Error sym/privBinaryOp (line 1032)

Csym = mupadmex(op,args{1}.s, args{2}.s, varargin{:});

Error sym/max (line 78)

C = privBinaryOp(X, Y, 'symobj::zipWithImplicitExpansion', 'symobj::maxComplex');

Error Gap (line 78)

tk=tk+max(-(d(T,seq(i),k)-a_halt_copy),0);

The only error is related to 'a_halt_copy'. Other part is ok. I really want to konw when I use 'max()' , how i can solve the function :

a_halt_copy-mu(seq(i))-sum(sum(sum(tk)))/K==0

I am a newer and learning knowledge about matlab. I beg for a help. This question has been bothering me for a long time. Thank you very much.

Answer 1

[Rahul]

Hi,

I understand that you’re trying to evaluate and solve symbolic expressions involving some variables, while applying “max” function in these expressions, although there are some unknown states in your provided code snippet:

• X_IJ: Function definition doesn’t exist leading to ambiguous definitions of ‘xij’ and ‘seq’ vectors. Although, we shall assume them to be a sample vector data later, wherever needed.
• bar_fix: Function isn’t defined leading to invalid assignment to variables ‘a_talt’, ’ref_bar_fix’.
• Heuristic: Function definition doesn’t exist, affecting the data in ‘a_opt’, ‘ref_heuristic’.

While solving the final expression in the first loop, for ‘a_halt_copy’ variable gives an indeterminate solution or an empty symbolic variable, which observed using the given code:

T=1;
N=10;
K=1000;
bar_W=20;
d=zeros(T,N,K);

% Assuming seq to be a ones vector
seq = ones(1, N - 1);   
a_halt=zeros(N-1,1);    

for i=1:N-1
    syms a_halt_copy   
    tk=0;
    for k=1:K    
        if i==1
            tk=tk+max(-(d(T,seq(i),k)-a_halt_copy),0);
        else
            tk=tk+max(-(bar_W+d(T,seq(i),k)-a_halt_copy),0);  
        end
    end    
    % Display symbolic expression      
    tk     
    exp = a_halt_copy-mu(seq(i))-sum(sum(sum(tk)))/k;    
    % Plot given expression over a_halt_copy in [60, 90]  
    fplot(exp, [60 90]) 
    eqn = exp==0;
    a_halt_copy=solve(eqn,a_halt_copy)    
    a_halt(i)=a_halt_copy;  
end

Warning: Unable to find explicit solution. For options, see help.

 
a_halt_copy =
 
Empty sym: 0-by-1
 

Unable to perform assignment because the left and right sides have a different number of elements.

This happens due to the equation having no explicit solution since the equation doesn’t exactly converge to 0, which leads to an empty symbolic variable being assigned to a vector element of ‘a_halt’. This can be observed by plotting LHS of equation using “fplot” (ran in R2024a):

Moreover, the usage of “sum” function in the given code, doesn’t provide any functionality since ‘tk’ is already a symbolic expression that can’t be summed.

Instead, you can try using “vpasolve” function to calculate solution with variable arithmetic precision as mentioned in this MATLAB answer:

https://www.mathworks.com/matlabcentral/answers/1632110-unable-to-find-explicit-solution

For more information regarding usage of “solve” or “vpasolve” functions in MATLAB, refer to the documentation links mentioned below:

• https://www.mathworks.com/help/symbolic/sym.solve.html
• https://www.mathworks.com/help/symbolic/sym.vpasolve.html