# The Monte Carlo tool uses a double execution of the S-main: # - pass #1 uses the scenario input.py file to process the variables identified # for dispersion. A specified number, N, of values {v_1, ..., v_N} is # generated for each variable v, with the values constrained by the specified # distribution of v; N is specified in the input file. # A set of N files, {RUN_1/monte_input.py, ... , RUN_N/monte_input.py} is # created, with each file containing one of the set of values for each # variable. Once these files are generated, the simulation is complete for # pass #1 and it terminates. # - pass #2 uses one of the generated files (monte_input.py) as the input file # for a regular execution of the simulation. There will typically be many # executions of the sim, one for each of the generated monte_input.py files. # This input file provides one example of how to test this two-pass process, # although it is admittedly a bit convoluted and hard to read. TODO: Once # TrickOps is capable of operating with this monte-carlo implementation, that # framework can manage both the generation and local execution of generated # monte_input.py files, removing the need for this type of "sim that launches a # sim" test methodology -Jordan 10/2022 # For the purpose of expedient testing, we generate and run only 2 files. # This is sufficient to demonstrate "multiple" without unnecessarily # burning CPU time. import os exename = "S_main_" + os.getenv("TRICK_HOST_CPU") + ".exe" # Pass #1 Generate the set of scenarios with unique dispersions print("Processing Pass #1 for run RUN_nominal") input_file = "RUN_nominal/input_a.py" ret = os.system("./" + exename + " " + input_file) if ret != 0: trick.exec_terminate_with_return(1, "double_pass.py", 34, "Error running " + input_file) # Pass #2 Run the scenarios. Logged data will go into each scenario's folder print("") print("") print("Processing Pass #2 for run RUN_nominal") for ii in range(2): input_file = "MONTE_RUN_nominal"+"/RUN_00%d/monte_input_a.py" %ii print ("**************** %s" %input_file) ret = os.system("./" + exename + " " + input_file) if ret != 0: trick.exec_terminate_with_return(1, "double_pass.py", 43, "Error running " + input_file) # To be compatible with our current unit-sim framework, this file has to be a # simulation input file. Therefore it needs a stop time so it doesn't run forever. trick.stop(0.0)