{"id":483,"date":"2019-03-28T21:25:17","date_gmt":"2019-03-28T21:25:17","guid":{"rendered":"http:\/\/ajhassoc.com\/?p=483"},"modified":"2019-03-29T02:03:46","modified_gmt":"2019-03-29T02:03:46","slug":"comments-on-forecasting-the-impossible","status":"publish","type":"post","link":"https:\/\/ajhassoc.com\/index.php\/2019\/03\/28\/comments-on-forecasting-the-impossible\/","title":{"rendered":"Comments on \u201cForecasting the Impossible\u201d"},"content":{"rendered":"
\n\n\n

Norman Marshall\u2019s article, \u201cForecasting the impossible: The status quo of estimating traffic flows with static traffic assignment and the future of dynamic traffic assignment,\u201d slams conventional modeling practices.\u00a0 Marshall states that MPOs routinely test alternatives that create unrealistically large flows on freeways and the consequential delays are so erroneous as to invalidate any impact analysis that uses them.\u00a0 The culprit is static traffic assignment (STA), and the fix is dynamic traffic assignment (DTA).<\/p>\n\n\n\n

I have been a strong advocate of DTA in planning studies of all types.  I do recognize that DTA is highly computational and that some larger MPOs might be better off putting their resources elsewhere.  In Ohio, several small MPOs have used area-wide DTA for their modeling for years by now, with excellent results.  NCHRP 765 notes examples of this and incorporates DTA into the \u201ctravel forecasting model ideal\u201d for project-level work.<\/p>\n\n\n\n

I am quite familiar with alternatives that assign traffic\nwell beyond a volume-to-capacity ratios (v\/c) of 1, and I cannot fathom why\nanybody would take any of this seriously, either as a realistic representation\nof the future or as a strawman case study. \nWhen a lot of facilities are badly over-assigned, then the model is\nbroken somewhere.  The failure may not be\nin the traffic assignment step but may be in the land-use allocation step or\nelsewhere.  Ideally, any alternative\nshould have all important facilities operating somewhere close to their design\ncapacity.  If this were the case, STA\nwould suffice for most planning.<\/p>\n\n\n\n

The problem largely stems from inaccurate estimates of delay\nfrom queuing in STA.  I have always\nbelieved that the funny shape of the BPR curve is intended to account for\nmoderate amounts of queuing, albeit rather roughly.  If 6000 cars try to squeeze through a\nfacility with a capacity of 4000 vehicles per hour, they will all eventually do\nit.  These vehicles will encounter a long\nqueue that will finally dissipate when volumes drop below capacity.  The BPR curve is not misguided, it is just a\npoor guess of queuing delay for many situations.  I will illustrate my point with a numerical\nexample.<\/p>\n\n\n\n

Consider a freeway with a lane drop causing a bottleneck, going from 3 lanes to 2 lanes.  The link upstream of the lane drop is very long.  There are four links, each 1-mile long downstream of the lane drop.  The capacity of each lane is 2000 vph and the free speed is 60 mph.  The upstream volume is 6000 vph during the peak hour.  Call this Case 1.  What is the downstream delay with STA or DTA?  I will adopt a BPR curve with \u201cstandard\u201d freeway coefficients as republished in NCHRP Report 365.  The free-speed multiplier (\u201calpha\u201d) is 0.83 and the v\/c exponent (“beta”) is 5.5.  (See Note 1)<\/p>\n\n\n\n

STA Case 1.  Each of\nthe 4 downstream links will have a v\/c ratio of 1.5.  A strict application of the BPR curve, as\nrequired by STA, gives a single-link travel time of about 8.7 minutes and a\n4-link time of 34.9 minutes.<\/p>\n\n\n\n

DTA Case 1.  DTA\ncalculates delay differently.  Each of\nthe 4 links will have a v\/c = 1 (approximately) and there will be an average\nqueuing delay of exactly 15 minutes. (see Note 2).  A downstream link travel time is 1.83\nminutes, so the four-link travel time is 7.3 minutes, and the total travel time\nis 22.3 minutes.<\/p>\n\n\n\n

STA overestimates delay.<\/p>\n\n\n\n

Now consider almost the same situation, but there is a work\nzone of 1 mile in length, thereby closing 1 lane of a 3-lane freeway for just\nthis one link.  Call this Case 2.  Otherwise, Case 2 is the same as Case 1.<\/p>\n\n\n\n

STA Case 2.  One of\nthe downstream links will have a v\/c of 1.5 and the other 3 links will have a\nv\/c of 1.0.  The total delay is 8.7 +\n3*1.83 = 14.9 minutes.<\/p>\n\n\n\n

DTA Case 2.  One link\nwill have a v\/c = 1 (see Note 3) and the three downstream-most links will have\na v\/c = 0.67.  Those further downstream\nlinks will have a travel time of 1.09 minutes, so the total average travel time\nis 3*1.09 + 1.83 + 15 = 20.1 minutes.<\/p>\n\n\n\n

STA now underestimates delay.<\/p>\n\n\n\n

So, you can see that STA and the BPR curve together create wildly inconsistent travel time estimates for over-capacity situations.<\/p>\n\n\n\n

STA will also tolerate over-capacity volume estimates when\nthe queuing penalty is not sufficiently large, even though there is plenty of\ncapacity on parallel routes.<\/p>\n\n\n\n

All these problems go away when there is ample capacity in\nthe system to handle all traffic.<\/p>\n\n\n\n

Here are Alan\u2019s guidelines for the use of STA in long range transportation planning.<\/p>\n\n\n\n