Troubleshooting Experiment 1: trying a long run, expanding boundaries further, and cross-checking parameters between article and code. Total time used: 10h 6m (25.2%)
Running one scenario with 100 population 50 generation 1 run on remote machine.
source ~/miniconda3/bin/activate
conda activate hernandez2015
python -m Experiment1Below, you can see the result from this for prescreen10 (top left). The others are all 100 population 5 generation from 1 run or 3 runs for prescreen50. As you can see, this unfortunately has no impact on the Y axis, and just the density of points.
Below is Figure 6 from prescreen10, just showing examples where forms > 6000.
I tried running with boundaries of 1-100 (rather than 1-60). Ran 100 population 5 generation 1 run.
Although this gets a more similar Y axis, the number of staff members became too high (so X axis becomes wrong), so that is not the solution.
Figure 6 (just showing forms > 10500)
I set the boundaries back to 1-60 (as that gave the correct range of staff members for Figures 5 and 6).
Looking at Figure 6 in the article, the number of line managers was always fairly low, but I find that to be as high as dispensers - although, in Table 4 (which is a sample of results) they likewise find high numbers of line managers.
I decided to go back over the paper, identifying all the model parameters mentioned, and make sure I could find these in the code and confirm they matched the paper.
I returned to this the following day, so please refer to subsequent logbook for results.
import sys
sys.path.append('../')
from timings import calculate_times
# Minutes used prior to today
used_to_date = 547
# Times from today
times = [
('09.15', '09.17'),
('11.40', '11.45'),
('12.07', '12.14'),
('15.54', '16.01'),
('16.02', '16.28'),
('16.42', '16.54')]
calculate_times(used_to_date, times)Time spent today: 59m, or 0h 59m
Total used to date: 606m, or 10h 6m
Time remaining: 1794m, or 29h 54m
Used 25.2% of 40 hours max