deployment - MATLAB deploytool simulink Undefined function 'load_system' -

i using matlab m file code loads simulink file, runs , evaluats many times inside complicated parallel loop.

it runs perfect @ matlab environment. when compiling code via deploytool command, installing , running command line, have difficult time.

at line loading model,

mymodel= ... ; load_system(mymodel); 

i following error

c:\program files\mymain\application>warning: name nonexistent or not directory: model > in path @ 109   in addpath @ 86   in run_main @ 3 warning: name nonexistent or not directory: data > in path @ 109   in addpath @ 86   in run_main @ 4 warning: name nonexistent or not directory: graphics > in path @ 109   in addpath @ 86   in run_main @ 5 starting parallel pool (parpool) using 'local' profile ... connected 2 workers. initializing ... undefined function 'load_system' input arguments of type 'char'.  error in run_main (line 40)   matlab:undefinedfunction 

i don't care warnings (as don't care of other matlab warnings) error not recognizing function open simulink model ('load_system') :

undefined function 'load_system' input arguments of type 'char'. 

i did search , found following questions:

load_system using mcr

deployment of simulink models

the first link says

i got confused: deployment tools compiling c++ or .net ? is supposed increase speed or becomming stand-alone advantage?

the problem both links talk strategy. how implement them? how can compile matlab code uses load_system , sim command load , simulate model?

can bring step step solution simple example of matlab code (m file) simulating model , compiling both of them?

the first sentence of documentation screen shot important one. mentioned @navan (in comment), simulink functionality not compatible matlab compiler. is, m-code uses simulink functinality (load_system, sim, etc) not execute when using matlab compiler.

assuming have access simulink coder, , model doesn't contain functionality prevents being converted c-code, option develop application in 2 parts:

1. with simulink model: use simulink coder create executable model. there several examples of doing in doc, including using rsim target batch simulations. there various command line options passing new parameter sets , inputs executable via mat file. results of simulation (i.e written matlab workspace) gets written mat file created executable.
2. within matlab code: replace simulink functionality call (typically using system function, shown in linked example) executable created in step 1. prior call you'll need create mat file parameters in it; , after call you'll want read results mat file created executable.

it's not trivial process, do-able.