rl-institut / simbev

MIT License
10 stars 4 forks source link

Missing EVs' CSVs in results #54

Closed nesnoj closed 2 years ago

nesnoj commented 2 years ago

I did some runs over the holidays (with latest release) and came across another problem. With this scenario the resulting number of EVs is too small (see sums at right of table):

region_id RegioStaR7 bev_mini bev_medium bev_luxury phev_mini phev_medium phev_luxury EXPECTED SUM ACTUAL SUM
71 SR_Metro 1690 3844 1235 1047 2382 765 10963 1371
72 SR_Gross 1750 3676 1045 1084 2278 647 10480 10480
73 SR_Mitte 3548 7776 2421 2197 4818 1499 22259 8536
74 SR_Klein 880 1958 574 545 1213 355 5525 3290
75 LR_Zentr 816 1760 511 506 1090 317 5000 1173
76 LR_Mitte 2050 4625 1331 1269 2866 824 12965 6886
77 LR_Klein 2250 5230 1487 1394 3241 921 14523 980

This seems to be a general problem. Any ideas, @timroepcke @Mosc5? Could this behaviour be related to the sampling?

nesnoj commented 2 years ago

Update:

Can you make sense of this?

mosc5 commented 2 years ago

When running the rs7 defaut scenario in a single thread, I get the following error:

grafik

Multithreading eats those kinds of errors and just continues with the next region, which explains why it hadn't been noticed before.