Optimisation with patternsearch solver evaluates objective function for same variable values twice

Hello everyone,

As the title states, I have an issue with patternsearch, it evaluates the same values for optimisation variables twice in a row. Solver options are default, except for FunctionTolerance, StepTolerance, InitialMeshSize (as it matters here, it's 5e-4, and poll has been successful so it's twice the size in iteration shown).

% Problem statement
problem.options = app.Optim_Options.Solver_Options; % optimoptitons   
problem.solver = app.Optim_Options.Solver_Choice; % patternsearch
problem.objective = @(x)Fun_Optim(x); 
problem.x0 = [app.Coordinates(1) app.Coordinates(3:8) app.Coordinates(10)]; % X1, X2, Y2, X3, Y3, X4, Y4, Y5
problem.lb = app.Optimisation_Vars.lb.*problem.x0;
problem.ub = app.Optimisation_Vars.ub.*problem.x0;
problem.Aeq = [1 -1 0 0 0 0 0 0; 
               0  0 0 0 0 0 1 -1];
problem.beq = [0; 0];

% In Fun_Optim(x) if it matters
function fobj = Fun_Optim(x)
% code
x = [x(1), x(2), x(3), x(4), x(5), x(6), x(7), x(8)]; 
X = [x(1)          , x(2), x(4), x(6), Coordinates(9)];
Y = [Coordinates(2), x(3), x(5), x(7), x(8)          ];
disp([X;Y]);
fobj = someotherfun(x, arg1, arg2)
end

What I could see from function evaluations, is that solver evaluates at x+InitialMeshSize twice, and that is something that I do not prefer as it's time expensive. Why would this even happen? I do have equality constraints x(1) = x(2) and x(7) = x(8) as seen in Aeq and Beq, and these are respected. Rest of the solver options are default.

Function evaluation : 7
   0.042840000000000   0.042840000000000   0.036280000000000   0.021000000000000                   0
                   0   0.021710000000000   0.034440000000000   0.042000000000000   0.042000000000000
fobj =   0.001044300000000
Command window display
2           7      0.0010443         0.002     Successful Poll
Function evaluation: 8
   0.043840000000000   0.043840000000000   0.036280000000000   0.021000000000000                   0 % Notice the first two numbers
                   0   0.021710000000000   0.034440000000000   0.042000000000000   0.042000000000000
fobj =   0.001474400000000
Function evaluation: 9
   0.043840000000000   0.043840000000000   0.036280000000000   0.021000000000000                   0 % Numbers did not change
                   0   0.021710000000000   0.034440000000000   0.042000000000000   0.042000000000000
fobj =   0.001474400000000
Function evaluation: 10
   0.042840000000000   0.042840000000000   0.036280000000000   0.021000000000000                   0 
                   0   0.021710000000000   0.034440000000000   0.043000000000000   0.043000000000000 % Last two Numbers changed
fobj =   0.001755000000000
Function evaluation: 11
   0.042840000000000   0.042840000000000   0.036280000000000   0.021000000000000                   0 
                   0   0.021710000000000   0.034440000000000   0.043000000000000   0.043000000000000 % Numbers did not change
fobj =   0.001755000000000
Function evaluation : 12
   0.041840000000000   0.041840000000000   0.036280000000000   0.021000000000000                   0
                   0   0.021710000000000   0.034440000000000   0.042000000000000   0.042000000000000
fobj =   7.588000000000008e-04
    3          12      0.0007588         0.004     Successful Poll
Stop requested.

I am curious why is this happening? I also ran the same problem without equality constraints and it seems that such issues do not exist.