Error using standalone function with ga toolbox

I have the following error message when running a standalone function with the genetic algortihm ga from the Global Optimization Toolbox:

Error using warning
Unable to load a message catalog 'globaloptim:gaminlpvalidateoptions'. Please check the file location and format.

Error in gaminlpvalidateoptions

Error in ga (line 295)

Error in test_GAopt (line 26)

MATLAB:builtins:MessageCatalogLoadFailed

I want to apply the genetic algorithm on a shared linux machine that doesn't have the Global Optimization Toolbox. For this I have created a virtual machine on my Windows computer from which I compile a standalone function that makes use of the genetic algorithm. The virtual machine has the same version of linux (scientific linux 6.9, 64-bit), and mcr (9.0). The standalone function runs succesful on the virtual machine.

The above error comes from a test function that roughly does what I desire:

function test_GAopt( N )

if isdeployed 
    N             = str2double( N );
end

%%
idx_real = 1:N/2;
idx_int  = (N/2+1):N;

fun     = @(x) x(idx_real)*x( idx_int )' + 2;

A_real          = cat( 1, -eye( N, 2*N ) ); 
A_int           = -[zeros( 1, N ) ones( 1, N )];  
A_ga            = cat( 1, A_real, A_int );
b_ga            = -[zeros( N, 1 ); 1];
lb_ga           = zeros( 1, 2*N );
ub_ga           = [Inf*ones( 1, N ) ones( 1, N )];

opt_ga      = gaoptimset( @ga );
opt_ga      = gaoptimset( opt_ga, 'Display', 'iter' );

[xsol,cost,exitflag_ga,output_ga] = ga( ... genetic algorithm
    fun, ...                           optimized function
    2*N, ...                                   number of variables
    A_ga, b_ga,...                              inequality constrains
    [],[], ...
    lb_ga, ub_ga, [], ...                       lower/upper bound
    idx_int, ...                                 index of integer variables
    opt_ga );

end

I compile the code using:

mcc('-m','test_GAopt.m', '-vm', '-R', '-logfile,log_GA_opt.txt', '-R','-nojvm','-R','-singleCompThread','-R','-nosplash', '-a', 'ga.m', '-a', '/usr/local/MATLAB/R2015b/toolbox/globaloptim', '-a', '/usr/local/MATLAB/R2015b/resources/globaloptim')

I included the -a options to try to solve the problem but also tried it without these options. I move the standalone file to the shared linux machine and make it executable.

Exucuting on the shared linux machine

./test_GAopt 100

gives the initial error. I suspect I have a problem with the paths, but I haven't been able to figure it out. Does anyone have the solution?

UPDATE

I found a workaround by making a copy of the Global Optimization toolbox directory, commenting out line 295 from ga.m, and compiling:

mcc('-m','test_GAopt.m', '-vm', '-R', '-logfile,log_GA_opt.txt', '-R','-nojvm','-R','-singleCompThread','-R','-nosplash', '-a', 'ga.m', '-a', './resources/globaloptim','-a', './toolbox/globaloptim/globaloptim/private')