Cube Voyager
swh220
—
2017-06-23T21:01:21Z —
#1
Dear Citilabs community,
My last posts on this forum sadly didn't receive any reactions
. This is my final try since i really need some results out of my model on which i worked pretty hard for a couple of weeks (simply learning the software took a lot of time). The research i'm doing is if the dutch car travelers can get an Pareto improvement from the introduction of dedicated autonomous vehicles (AV) lanes. That is if AV's are penetrating the market is can travel cost of both AV users and regular car users decrease by implementing such kind of lanes(at least one user group gains while non lose). Now i have aggregated my OD matrices provided from the Dutch road provider as well as my Zones and network. I developed a script for which i thought the program could provide me reasonable multiclass assignment results. Now the base scenario (no AV vehicles) gives me unrealistic results and im not sure wheter or not the linkread phase reads my NETI file correctly. I think im close to good results but feel like i just can't make te final step. Help is more than welcome. I will attach my script and network file and my OD matrix and print file with strange wrong results. The solution is probably not far away. If you guys maybe now an online CUBE teacher or an offline one in Holland i love to hear it and willing to pay a reasonable amount of money to get some results out of my model.
InitialAssignment_0000.PRN (1.0 KB) AVassignmentscript.S (2.3 KB) Randstadnetwerkgoedeagzonenumberscentroidnumberszonderoverbodige.net (667.9 KB) INGEKROMPEN MATRIXaverage.MAT (33.2 KB)
Thanks in advance
clanbaby
—
2017-06-24T00:24:16Z —
#2
.....*.s script file can't download from the Community! plz Check.
If you want to Pareto Improvment, it looks like in the assignment can not use UE,but SO or an improved SO.
So ,in your last topic, the script "PARAMETERS COMBINE=EQUI" maybe something wrong.
swh220
—
2017-06-25T20:17:35Z —
#3
Thanks for your effort to respond! I will post my script under this message. I was presuming that the PARAMETER EQUI command give me the nash equilibrium traffic flows so that no user have an incentive to change their path, because otherwise their travel cost will increase. This can mean that the network is in equilibrium but does not neccesarly mean that the equilibrium is socially optimal (like in the classic prisoners dilemma) especially in a static assginment. A road provider might allocate traffic flows otherwise if they could, by for example road pricing (i ignore this in my study because road pricing is still not accepted in the Netherlands). Im not exactly sure what you mean by the assignment can't use UE. Is this their something wrong with my script (i changed it since my last post though) or with the command? I like the Frank-Wolfe algorthim and therefore chose the EQUI command. I'm happy to hear suggestions to improve my method and thus, my chosen algortihm to converge.
RUN PGM=HIGHWAY PRNFILE="C:\EX_MC_EQUI\MCHWY00J.PRN" MSG='LW.COSTs & FUNC COST'
FILEI MATI[1] = "C:\EX_MC_EQUI\MCMAT00A.MAT"
FILEO NETO = "C:\EX_MC_EQUI\MCHWY00J.NET"
FILEI NETI = "C:\EX_MC_EQUI\data\BASE.NET"
DIST_COST_HV = 0 ; HV Normalized to 0
DIST_COST_AV = {AV_DISTANCE_COST} ; Additional AV distance cost per meter
HV_VOT=0.15 ;VOT per minute
AV_VOT=HV_VOT*{AV_VOT_Proportion} ; VOT AV per minute 80% of regular
PAR COMBINE=EQUI MAXITERS=100 GAP=0.0001
PROCESS PHASE=LINKREAD
IF (LI.LINKTYPE ==3) C=LI.H_CAP*{AV_LINK_CAPAC} ; Extra capacity AV link (2*2.5)
ELSE
C=LI.H_CAP*2 ;capacity from input network from times 2 (2 hour nightly rush hour)
DISTANCE=LI.AFSTAND ; Afstand is Length of links from input network file in meters
IF(A<134 || B<134) ;These are the centroid connectors created by cube
T0=0.001
ELSE
TO=((LI.AFSTAND/1000)/(100/60)) ;FREE flow time in minuten Veranderen naar wettelijke snelheid ipv 100 km/h
ENDIF
LINKCLASS=LI.LINKTYPE ; LINKTYPE is obtained from input file
IF (LI.LINKTYPE == 1) LINKCLASS=1 ; Autosnelweg
IF (LI.LINKTYPE == 2) LINKCLASS=2 ; Autoweg
IF (LI.LINKTYPE ==3) LINKCLASS=3 ; AV link
IF (LI.LINKTYPE=0-2) ADDTOGROUP=1 ; GROUP 1 is HV users
IF (LI.LINKTYPE=3) ADDTOGROUP=2 ; GROUP 2 is AV users (linktype AV lane has to be set to 3
IF (ITERATION=0)
LW.COST_HV = T0 + DIST_COST_HV * LI.AFSTAND
LW.COST_AV = T0 + DIST_COST_AV * LI.AFSTAND
ENDIF
ENDPROCESS
PROCESS PHASE=ILOOP
PATHLOAD PATH = LW.COST_HV VOL[1]=MI.1.1*{HV_share_demand} EXCLUDEGROUP=2 ; HV vehicles cant use AV paths that contain AV lanes
PATHLOAD PATH = LW.COST_AV VOL[2]=MI.1.1*{AV_share_demand} ; AV demand
ENDPROCESS
PROCESS PHASE=ADJUST
FUNCTION {
V=VOL[1]+VOL[2]
TC[0] = TO ; Centroid connector
TC[1] = T0 * (1+0.2*(V/C)^10) ; Updated BPR funtion freeway
TC[2] = T0 * (1+0.05*(V/C)^10) ; Updated BPR function arterial
}
LW.COST_HV = TIME*HV_VOT + DIST_COST_LOV * LI.AFSTAND
LW.COST_AV = TIME*AV_VOT + DIST_COST_HOV * LI.AFSTAND
FUNCTION COST= ((TIME + DIST_COST_HV * LI.AFSTAND) * V1 + (TIME + DIST_COST_AV * LI.AFSTAND) * V2) / CmpNumRetNum(V,'=',0,1,V) ; cost function
ENDPROCESS
PROCESS PHASE=CONVERGE
ENDPROCESS
ENDRUN
Filippo
—
2017-06-26T08:49:36Z —
#4
I noticed a few errors and sent you an e-mail through the support channel.
Oliver
—
2017-06-26T09:32:11Z —
#5
Hello,
Citilabs online courses can be found here: http://citilabs.teachable.com/courses
Please contact me if you would like more information: ocharlesworth@citilabs.com
Thank you
Oliver
Asad
—
2018-07-11T21:24:55Z —
#6
Hi swh220,
Can I contact you inbox? I am working on AV assignment.
Thanks,
Asad.