Example 8: Emax Model, PSO

Example 8 uses the PSO algorithm on a simulated dataset using an Emax model with Body Weight (BW) having a power effect on E0, Sex and Race having an exponential effect on E0, sigmodicity = 1, and a proportional error model. Additionally, a final downhill search is performed.

The template file can be downloaded here and the tokens file here.

The options file looks like:

{
  "author": "Certara",
  "algorithm": "PSO",
  "PSO": {
    "cognitive": 0.5,
    "social": 0.9,
    "inertia": 0.9,
    "neighbor_num": 2,
    "p_norm": 1
  },
  "num_parallel": 4,
  "random_seed": 11,
  "population_size": 10,
  "num_generations": 7,
  "downhill_period": -1,
  "num_niches": 2,
  "niche_radius": 2,
  "local_2_bit_search": false,
  "final_downhill_search": true,
  "penalty": {
    "theta": 5,
    "omega": 5,
    "sigma": 5,
    "convergence": 100,
    "covariance": 100,
    "correlation": 100,
    "condition_number": 100,
    "non_influential_tokens": 0.00001
  },
  "crash_value": 99999999999,
  "remove_run_dir": false,
  "nmfe_path": "c:/nm74g64/util/nmfe74.bat",
  "model_run_timeout": 1200
}

and can be downloaded here.

Execute Search

Usage details for starting a search in pyDarwin can be found here.

See “Examples” for additional details about accessing example files.

Initialization output should look like:

[12:56:38] Model run priority is below_normal
[12:56:38] Using darwin.MemoryModelCache
[12:56:38] Algorithm is PSO
[12:56:38] Project dir: C:\Workspace\Example8
[12:56:38] Data dir: C:\Workspace\Example8
[12:56:38] Project working dir: C:\Users\jcraig\pydarwin\Example8
[12:56:38] Project temp dir: C:\Users\jcraig\pydarwin\Example8\temp
[12:56:38] Project output dir: C:\Users\jcraig\pydarwin\Example8\output
[12:56:38] Writing intermediate output to C:\Users\jcraig\pydarwin\Example8\output\results.csv
[12:56:38] Models will be saved in C:\Users\jcraig\pydarwin\Example8\models.json
[12:56:38] Template file found at template.txt
[12:56:38] Tokens file found at tokens.json
[12:56:38] Search start time = Fri Dec 16 12:56:38 2022
[12:56:38] NMFE found: c:/nm74g64/util/nmfe74.bat
[12:56:38] Not using Post Run R code
[12:56:38] Not using Post Run Python code
[12:56:38] Checking files in C:\Users\jcraig\pydarwin\Example8\temp\0\02
[12:56:38] Data set # 1 was found: C:\Workspace\Example8\PDdata.csv
[12:57:03] Iteration = 0, Model     2,           Done,    fitness = -619.152,    message = No important warnings
[12:57:05] Iteration = 0, Model     4,           Done,    fitness = -2475.044,    message = No important warnings
[12:57:11] Iteration = 0, Model     3,           Done,    fitness = -723.686,    message = No important warnings

and the final output should look like:

[13:13:01] Iteration = 6D03, Model    15,           Done,    fitness = -2822.644,    message = No important warnings
[13:13:05] Iteration = 6D03, Model    16,           Done,    fitness = -2821.427,    message = No important warnings
[13:13:14] Iteration = 6D03, Model    20,           Done,    fitness = -2820.297,    message = No important warnings
[13:13:14] best fitness -2829.911515459548, model [1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1]
[13:13:14] Final output from best model is in C:\Users\jcraig\pydarwin\Example8\output\FinalResultFile.lst
[13:13:14] Number of unique models to best model = 72
[13:13:14] Time to best model = 13.4 minutes
[13:13:14] Best overall fitness = -2829.911515, iteration 6D02, model 6
[13:13:14] Elapsed time = 16.6 minutes