Is any way, how to distribute intrazonal trips not only with fixed time, but widely i.e. as a interval (0,28) or "<=28".Background: Average trip duration is aprox 28 minutes (sociodem survey) and average intrazonal cost is in time is also 28 minutes. Standard computing of intrazonal trip in distribution use discrete number 28. I want use use 27,26,25 ... to 0 min.classic discrete distribution code: SETPA P[1]= ZI.1.P1 A[1]= ZI.1.A1 GRAVITY PURPOSE=1, LOS=MW[40], FFACTORS=FF
Hi Daniel.
I am not sure if I fully understand your question, so perhaps I need a clarification. I think that you are saying that your distribution model use a fixed intrazonal time = 28 minutes. My understanding is that you would like to have a random value between 0 and 28 instead of the fixed value for intrazonal values.
If this is the case, you could use something like MW[#][I] = RANDOM(29).
Alternatively, you can use other ways of calculating the intrazonal times, like the "half-time-to-the-nearest-neighbor" method, that could also be easily implemented in CUBE.
Please let me know.
Dear Filippo, thanks for your answer.My problem is not how to calculate INTRAZONAL discrete time (or generate variability in diagonal cells), but how to work with intrazonal time in distribution as a interval for each one zone.My model is mixed area with a small and large municipalities. Many small village with 100 citizen have INRAZONAL time value 3 minutes and gravity distribution work well, because this time represented marginal trips (by frequency function). We can neglected in reality, that trips have duration 1, 2 or 3 minutes .But town, with 100 000 citizen is large area and INRAZONAL time value is 28 minutes. The 28 minutes mean, that INRAZONAL trips are 28,27,26,25,...,2,1 minutes in reality. But distribution work with the time as fixed value - 28 minutes and do not created INTRAZONAL trips with time below 27 minutes.FrequencyFunction: ; time in min, purpose no. 1 1.01 06 0.25916 0.36826 0.17836 0.09846 0.05756 0.01166 0.01176 0.00686 0.01196 0--------------------------------Frequency OD trip duration in minutes - What I need:--------------------------------------- 1 - 11 43 206.77 6.5 6.5 |====== 11 - 21 99 1,478.18 46.4 52.8 |============================================== 21 - 31 69 936.44 29.4 82.2 |============================= 31 - 41 39 414.56 13.0 95.2 |============ 41 - 51 18 149.29 4.7 99.9 |==== 51 - 61 2 3.76 0.1 100.0 | 61 - 71 -- -- -- 100.0 | 71 - 81 -- -- -- 100.0 | 81 - 91 -- -- -- 100.0 | 91 - 100 -- -- -- 100.0 | 100+ -- -- -- 100.0 |-------------------------------------------------------------------Frequency OD trip duration in minutes - What is result:--------------------------------------- 1 - 31 567 18,052.47 0.5 0.5 | 31 - 61 1,917 584,580.21 16.3 16.9 |================ 61 - 91 9,492 21,317.14 0.6 17.4 | 91 - 121 14,391 17,447.82 0.5 17.9 | 121 - 151 16,989 281,617.17 7.9 25.8 |======= 151 - 181 21,176 324,330.81 9.1 34.9 |========= 181 - 211 23,112 18,912.22 0.5 35.4 | 211 - 241 20,383 23,538.28 0.7 36.1 | 241 - 271 14,302 756,169.00 21.1 57.2 |===================== 271 - 301 8,258 260,373.44 7.3 64.5 |======= 301 - 331 5,208 830,919.50 23.2 87.7 |======================= 331 - 361 4,052 54,155.63 1.5 89.2 |= 361 - 391 3,000 4,864.23 0.1 89.4 | 391 - 421 2,465 15,520.76 0.4 89.8 | 421 - 451 1,926 364,370.33 10.2 100.0 |========== 451 - 481 -- -- -- 100.0 | 481 - 500 -- -- -- 100.0 | 500+ -- -- -- 100.0 |----------------------------------------
Hello.
From your clarification comment my understanding is that if 28 minutes is the intrazonal average for a specific zone, this means that the actual intrazonal trips that you have can have values above or below 28, and that 28 is the mean, and not that you can only have values <= 28 minutes. Please let me know if I am misunderstanding.
Distribution models like the gravity model calculates trip distribution between zones i-j, e.g., 1-1, 1-2, 1-3, etc. based on the impedance within/between zones, and the value of the intrazonal impedance is important to avoid biased estimates.
If my understanding of your problem is correct, I think that the same considerations that you are describing for intrazonal impedances applies also for interzonal impedances.
My understanding is that it might be possible that you might need to refine your zoning system due to the large sizes of these zones.
HI! Yes, you wrote it more accurately.INTRAZONAL SKIM value is average value, but in really is a interval of values, probably with different weight.If this value is low, i.g. 2 minutes, and zone is small, than variance of values is insignificant.If value is higher, i.g. 28 minutes, and zone is with huge production/attraction, than variance of value is significant. We need more accurate solution. You wrote, that is possible replace one big zone with more zones and reduce INTRAZONAL SKIM value. Replacing of one zone is possible with:a) real zones, if we have data (i have not it)b) fictive zones connected to one node, each one zone represented other value of intrazonal SKIM time (for 28 e.g. 8,13,18,23,28,33,38,43,48), same extrazonal SKIM time and partial value of production/attraction.But why we do by hand, if we have best modelling sw?
What you are describing is not a common practice. The common practice is to break large zones into appropriate sized zones to calibrate the model to the desired trip distribution. Our Voyager scripting is pretty flexible enough to implement most conditions and you should be able to set some thing like what you are describing. But it will not be a straightforward implementation.