strtok function takes linear optimization into an infinite loop

Question

Edoardo Beduglio 2024-11-12，16:13

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2165669-strtok-function-takes-linear-optimization-into-an-infinite-loop

评论： Swastik Sarkar 2024-11-20，4:13

I wrote a linear optimization algorithm to find the optimal dispatch of a battery system:

Xdam_d = optimvar('Xdam_d',T,N,'LowerBound',0,'UpperBound', Xbatt,'Type','continuous');
Xdam_c = optimvar('Xdam_c',T,N,'LowerBound',0,'UpperBound', Xbatt,'Type','continuous');
Xidm_d = optimvar('Xidm_d',T,N,'LowerBound',0,'UpperBound', Xbatt,'Type','continuous');
Xidm_c = optimvar('Xidm_c',T,N,'LowerBound',0,'UpperBound', Xbatt,'Type','continuous');
SOC = optimvar('SOC',T+1,N,'LowerBound',SOCbatt_min,'UpperBound',SOCbatt_max,'Type','continuous');

chargecons1 = Xdam_c + Xidm_c <= Xbatt*u_c;
dischargecons1 = Xdam_d + Xidm_d <= Xbatt*(1 - u_c);
Pch_eq = Pch == Xdam_c + Xidm_c;% - Xfrc_ind_a_up;
Pdis_eq = Pdis ==  Xdam_d + Xidm_d;% - Xfrc_ind_a_down;
for d = 1:N
    % loop on quarters of hours (t)
    for t = 1:T
        % Define the Income (R-C) for the battery on the multi-markets
        Idam(t,d) = pdam(t,d).*(Xdam_d(t,d) - Xdam_c(t,d))*0.25; %Day_Ahead market
        Iidm(t,d) = pidm(t,d).*(Xidm_d(t,d) - Xidm_c(t,d))*0.25; %Intraday
    end
   maxProfit = sum(Idam,'all') + sum(Iidm,'all');
   prob.Objective = maxProfit;

Then I call the constraints and the solver..

but the problem came when I added equation State of Charge, defined as a constraint and shown below :

soc_constraints(1,1) = SOC(1,1) == SOC_0;
soc_constraints(2:T+1,:) = SOC(2:T+1,:) == SOC(1:T,:) + (Pch(1:T,:)*etabatt_ch-Pdis(1:T,:)/etabatt_dis)*(100/4*Ebatt);

If I run the algorithm, it does not converge, or rather it enters a loop that takes it to MaxTime. If I pause, I can see that the solver stops on the strtok function and particularly at this point:

if nargin < 1 || nargin > 2

Whereas if I remove one of the two contributions Xdam or Xidm from the SOC function (especially at the discharge side) the solver finds optimal solution.

Or similarly if I remove the percentage factor (*100) it works, but obviously the SOC calculation is not correct.

Where could the problem be?