Using the example `treat-sim` model

For examples in this section, we’ll use the treat-sim simulation model. This is an example model of a terminating system based on exercise 13 from page 170 of:

Nelson. B.L. (2013). Foundations and methods of stochastic simulation. Springer.

Installing the model

The model is available on GitHub, and can be installed from PyPI:

pip install treat-sim

# pylint: disable=missing-module-docstring
from treat_sim.model import Scenario, single_run, multiple_replications

Running a single replication

Create an instance of Scenario. This contains all the parameters used in the model (e.g. resources and statistical distributions for arrivals and activities).
Run the model once with single_run(). This function is passed an instance of Scenario and an integer to set the random numbers used.

This returns a pandas.DataFrame with all performance measures for that replication, e.g. waiting times, utilisation, throughput.

scenario = Scenario()

# Run a single replication
results = single_run(scenario, random_no_set=1)
print(results.T)

rep                                      1
00_arrivals                     227.000000
01a_triage_wait                  57.120114
01b_triage_util                   0.621348
02a_registration_wait            90.002385
02b_registration_util             0.836685
03a_examination_wait             14.688492
03b_examination_util              0.847295
04a_treatment_wait(non_trauma)  120.245474
04b_treatment_util(non_trauma)    0.912127
05_total_time(non-trauma)       233.882040
06a_trauma_wait                 133.813901
06b_trauma_util                   0.834124
07a_treatment_wait(trauma)        3.715446
07b_treatment_util(trauma)        0.367507
08_total_time(trauma)           301.521195
09_throughput                   161.000000

You can extract a single metric like this:

print(results["01a_triage_wait"].iloc[0])

57.120114285877285

Running multiple replications

Running several replications at once is very similar to running a single replication — you just use the multiple_replications() function instead. You’ll need to provide an instance of Scenario, along with the number of replications you want to run.

scenario = Scenario()

# Run multiple replications
results = multiple_replications(scenario, n_reps=20)
results.head()

	00_arrivals	01a_triage_wait	01b_triage_util	02a_registration_wait	02b_registration_util	03a_examination_wait	03b_examination_util	04a_treatment_wait(non_trauma)	04b_treatment_util(non_trauma)	05_total_time(non-trauma)	06a_trauma_wait	06b_trauma_util	07a_treatment_wait(trauma)	07b_treatment_util(trauma)	08_total_time(trauma)	09_throughput
rep
1	230.0	24.280943	0.613250	103.242292	0.854504	31.089680	0.861719	152.483394	0.890904	234.759918	236.508444	1.028887	13.701783	0.607996	346.698079	171.0
2	227.0	57.120114	0.621348	90.002385	0.836685	14.688492	0.847295	120.245474	0.912127	233.882040	133.813901	0.834124	3.715446	0.367507	301.521195	161.0
3	229.0	28.659383	0.573698	112.242503	0.848514	21.374092	0.856306	94.019885	0.868888	208.361290	276.422566	0.874245	12.252175	0.464740	440.515502	167.0
4	240.0	24.801392	0.610618	121.535491	0.867779	31.556444	0.868617	150.351654	0.890564	249.690045	149.374881	0.879746	16.520898	0.663415	281.060924	172.0
5	233.0	17.679156	0.643085	103.606078	0.870109	25.505255	0.882532	157.255345	0.925467	233.296674	250.712956	0.862277	9.225844	0.421116	380.832994	164.0