scrit can't error related to Sortrows and matrix

I tried to run script below, but always have this error linked to sortrows, I checked sol matrice size and the number of column, but I don't understand the issue. If I hide the sol line the script will run, but with wrong results. Please any help to solve this issue?

Error using matlab.internal.math.sortrowsParseInputs Column sorting vector must contain integers with absolute value between 1 and the number of columns in the first argument. Error in sortrows (line 64) [col, nanflag, compareflag] = matlab.internal.math.sortrowsParseInputs(ismatrix(A),size(A,2),A,varargin{:}); Error in TLP (line 110) sol_sorted = sortrows(sol,8);

% Initialisation
clc
clear
format short g
format compact
% ------------------------------------------------------------------------
pi = 3.141592654;               % 
rho_Water = 1.025;              % Water density
% Inputs Part 1: Design parameters Spar; Ballast; Truss cavle; Column Water height
% TOPSIDE PROPERTIES geometry %% Unit m %%
Length = 112;
Width =	58;
Height = 36;
MassT = 25000;
VCG = 16;
Rxx = 23;
Ryy = 45;
Rzz = 51;
% DECK PROPERTIES geometry %% Unit m %%
Deck_Length = 112;
Deck_Width = 58;
Deck_Clearance = 18;
% Tendons PROPERTIES geometry %% Unit m %%
Ltd = 400;
% MASS FACTOR	%% Unit T/m3 %%
% Region	Factor
Nm = 8;   % Number of mooring
Tm = 400; % Tendon lenght
% MASS FACTOR	%% Unit T/m3 %%
% Region	Factor		
rhoNd = 0.25;
rhoPT = 0.2;
rhoCL = 0.15;
rhoTD = 0.2;
% Metacentric height
MetaC = 1;
% Inputs Part 2: Search Parameters (for a 2D Exhaustive Search) ; all parameters in m
Dc1 = 14;                    % Lower Total spar draft
Dc2 = 30;                    % Upper Total spar draft
tDc = 2;                     % Grid size in Draft direction
Hc1 = 25;                     % Lower Ballast External height
Hc2 = 50;                     % Upper Ballast External height
tHc = 5;                      % Grid Ballast External height
Hp1 = 5;                    % Lower Ballast External Diameter
Hp2 = 11;                    % Upper Ballast External Diameter
tHp = 1;                    % Grid Ballast External Diameter // tDb_out = 1
Mt1 =  5000;                    % Lower Ballast Internal Diameter //4
Mt2 = 12000;                    % Upper Ballast Internal Diameter
tMt = 1000;                  % Grid Ballast Internal Diameter
DeckH1 = 4;                    % Lower Ballast Internal Diameter //4
DeckH2 = 16;                    % Upper Ballast Internal Diameter
tDeck = 2; 
Dtd1 = 0.1;                    % Lower Ballast Internal Diameter //4
Dtd2 = 0.8;                    % Upper Ballast Internal Diameter
tDtd = 0.1; 
Dt1 = 20;
Dt2 = 30;
tDt = 5;
% ------------------------------------------------------------------------
% Exhaustive Search
sol=[];						% Start with an empty matrix of feasible solutions
data = [];              % Start  []
% data_feasible = [];	% these are only used to show the search space and feasible/infeasible solutions
for Dc=Dc1:tDc:Dc2
    for Hc=Hc1:tHc:Hc2
        for Hp=Hp1:tHp:Hp2
            for Mt=Mt1:tMt:Mt2
                for Dtd=Dtd1:tDtd:Dtd2
                    for DeckH=DeckH1:tDeck:DeckH2
                        for Dt=Dt1:tDt:Dt2                             
                            data = [data; Dt, DeckH, Dtd, Mt, Hp, Hc, Dc]; %, Dts
                            Cs = Width - Dc;
                            Lp = Width - 2*Dc;
                            lpt = Hp;
                            Vc1 = pi*(Hc-Dt)*(Dc^2)/4;
                            Vc2 = pi*(Dt-Hp)*(Dc^2)/4;
                            Vn = pi*Hp*(Dc^2)/4;
                            Vp = Hp*lpt*Lp;
                            Vtd = pi*Tm*(Dtd^2)/4;
                            V = 4*(Vc2 + Vn + Vp) + Nm*Vtd ;
                            OB = (4*Vc2*(-(Dt-Hp)/2) + 4*Vn*(-(Dt-Hp)-Hp/2) + 4*Vp*(-(Dt-Hp)-Hp/2) + Nm*Vtd*(-(Dt-Hp)-Hp-Tm/2))/V ;
                            W = MassT + Mt + 4*(Vc1+Vc2)*rhoCL + 4*Vn*rhoNd + 4*Vp*rhoPT + Nm*Vtd*rhoTD;
                            OG = (((Hc-Dt)+DeckH+Height/2)*MassT + ((Hc-Dt)+DeckH/2)*Mt + (((Hc-Dt)/2)*Vc1*rhoCL*4) + ((-4*(Dt-Hp)/2)*Vc2*rhoCL) + ((-4*(Dt-Hp)+Hp/2)*Vn*rhoNd) + ((-4*(Dt-Hp)+Hp/2)*Vp*rhoPT) + ((-Nm*(Dt-Hp)+Hp+Tm/2)*Vtd*rhoTD))/W ;
                            GM = OB - OG + 4.*pi.*((Dc).^4./(64.*V));
                            Fb = W./0.75;
                            T = W./3;
                            To = T./Nm;
                            %                             Dt = 4*((Fb/(4*rho_Water))-(Width - 2*Dc)*Hp)/(pi*Dc.^2);
                            if (GM>MetaC)%&& Dc < Cs && Dc>lpt && Dt>Hp && Hc>Dt
                                %                             data_feasible=[data_feasible; Dc, Hc, Hp, Mt, Dtd, DeckH, Dt];
                                W = MassT + Mt + 4.*(Vc1+Vc2).*rhoCL + 4.*Vn.*rhoNd + 4.*Vp.*rhoPT + Nm.*Vtd.*rhoTD ;
                                %                           TMass = MassT + Mt + 4*(Vc1+Vc2)*rhoCL + 4*Vn*rhoNd + 4*Vp*rhoPT + Nm*Vtd*rhoTD ;
                                sol=[sol; Dt, DeckH, Dtd, Mt, Hp, Hc, Dc, W]; 		%#ok<AGROW> % Store the feasible solution
                            end
                        end                     
                    end
                end
            end 
        end
    end
end
% sort the matrix of feasible solutions based on the objective (here vol, the third column)
sol_sorted = sortrows(sol,8);
% Select the first solution (which is the optimum after sorting all solutions based on the objective)
opt_sol=sol_sorted(1,:);
% ------------------------------------------------------------

Answer 1

[Steven Lord]

% sort the matrix of feasible solutions based on the objective (here vol, the third column)
sol_sorted = sortrows(sol,8);

Does your sol matrix have at least 8 columns? From the error message it seems like the answer is no.

The fact that the comment refers to "the third column" makes that look even more suspicious.

[Rigo ZOO]

when checked the matrix sol size I have

sizeOfMatrix =

0 0

Why this matrix is zero?

[Bruno Luong]

I suspect the conditional test

if (GM>MetaC)%&& Dc < Cs && Dc>lpt && Dt>Hp && Hc>Dt

is never true so you actually never grow sol.

Learn to use debugger when develop codes.

Answer 2

[Bruno Luong]

Replace initialization statements

sol=[];						% Start with an empty matrix of feasible solutions
data = [];

with

sol=zeros(0,8);						% Start with an empty matrix of feasible solutions
data = zeros(0,7);  

[Bruno Luong]

The code eventually still crash but in the next line

opt_sol=sol_sorted(1,:);

[Rigo ZOO]

Hi,

I changed sol[] and data[] to sol=zeros(0,8) and data = zeros(0,7), but I have the error mentioned below

ans =

0 8

ans =

'double'

Index in position 1 exceeds array bounds.

Error in TLP (line 116)

opt_sol=sol_sorted(1,:);

[Torsten]

To repeat Bruno's answer:

I suspect the conditional test

if (GM>MetaC)%&& Dc < Cs && Dc>lpt && Dt>Hp && Hc>Dt

is never true so you actually never grow sol.

Learn to use debugger when develop codes.

The "sol" matrix remains empty in your code - thus also "sol_sorted(1,:)" does not exist.

[Bruno Luong]

@Rigo ZOObut I have the error mentioned below

Index in position 1 exceeds array bounds.

Please replace

opt_sol=sol_sorted(1,:); 

by

if isempty(sol_sorted)
    error('your exhaustive search is wrong and did not find a solution; you need to double check your code and the parameter values (MetaC)');
end
opt_sol=sol_sorted(1,:);