KYOVA Metropolitan Planning Organization

INTRODUCTION AND BACKGROUND

Travel demand modeling for Huntington Ironton Area Transportation Study (HIATS) was performed using the QRSII transportation planning computer software package. This software will enable KYOVA to validate transportation models and also to forecast volumes for future socioeconomic conditions.

The previous model developed in the mid-1990�s was utilized in developing the 2025 Long-Range Transportation Study for the Huntington Ashland-Ironton (HAITS) region by the KYOVA Interstate Planning Commission [1]. The previous travel demand model utilized QRSII software and represented the areas KYOVA was responsible for in 1990 (Cabell and Wayne Counties in West Virginia and Lawrence County in Ohio) and excluded the Kentucky Counties. The current model includes Cabell and Wayne Counties in West Virginia, Lawrence County in Ohio and Boyd and Greenup Counties in Kentucky.

The decision was made to utilize the framework of the HAITS model and supplement it with the recently completed Ashland Travel Model conduct by the Corradino Group [2]. The Ashland model was developed in TransCAD which is not directly compatible with the QRSII software. The parameters of the HAITS model were developed by earlier studies used to produce the 1990 and 2000 Long-Range Plan [3][4].

The major activities involved:

1. Expand the traffic analysis zone (TAZ) structure to include all of Cabell, Wayne and Lawrence Counties. The TAZ�s structure developed for the two Kentucky Counties, Boyd and Greenup Counties, as part of the Ashland Travel Model were adopted for inclusion in this study. The TAZ�s by county are as follows:

	Wilbur Smith - HAIATS Model [4}	HAITS Model [1]	2004 Updated Model
Base Year Data	1970	1990	2000
Lawrence Co., Ohio	88	88	98
Cabell Co., WV	138	138	158
Wayne Co., WV	41	41	47
Greenup Co. KY	130	--	60
Boyd Co., KY		--	120
Internal Zones � Total	397	267	483
External Stations - Total	28	28	27

In Ohio and West Virginia the expanded TAZ structure added an additional 57,379 persons over a base population of 143, 772 from the 1990 study (Table 1). The revised TAZ structure for the Ohio and West Virginia Counties was presented as the �1999 TAZ District Map� in the KYOVA Year 2025 Long Range Transportation Plan [1]. The 1990 socio-economic profile for each TAZ was updated from the 2000 census. The updated variables for each TAZ included:

a. Dwelling Units

b. Population

c. Labor Force

d. Non-Retail Employment

e. Retail Employment

f. Automobile Ownership

g. School Attendance

h. Resident Students

The network was extensively edited to remove extraneous nodes and links in order to better reflect the regional highway system. Additional roadways constructed since the original networks were added as links.

Table 1 Population By Counties Included in 2004 Travel Model

	TAZ Number	1990 Population	2000 Population
Lawrence Co., OH	1-88	52,732	48,585
	600-609	---	12,892
	Sub Total	52,732	61,477
Wayne Co. WV	186-226	19,042	19,446
	550-555	---	23,438
	Sub Total	19,042	42,884
Cabell Co. WV	89-185 227-267	85,123	75,741
	550-519	--	21,049
	Sub Total	85,123	96,790
WV and OH Study Area		156,897	201,151
Greenup Co. KY		---	36,891
Boyd Co., KY		---	49,752
Total Study Area		156,897	287,794

Source: [1], Chpt. 6

Nodes-Point Data:

The nodes represent: 1) intersections that facilitate traffic flow between links or between traffic analysis zone centroids and the network, 2) external stations and 3) �dummy� nodes to more accurately reflect curvature along the road. In the model there are:

� external stations - 27

� zone centriods- 483

� nodes � 7300

� signalized intersections � 223 (20 in Wayne County, 35 in Lawrence County, 147 in Cabell County and 38 in Boyd/Greenup Counties).

Zone centroids represent the trip ends associated with the origin and destination of trips entering the roadway networks. External stations represent the origin and/or destination of trips having at least one end located externally to the study area. The data included for each node representation are as follows:

BACKGROUND AND OVERVIEW

1. A BRIEF DESCRIPTION OF

THE QUICK RESPONSE SYSTEM II ( QRSII)

The Quick Response System II is a computer program for forecasting impacts of urban development on the highway traffic and for forecasting impacts of highway projects on travel patterns.

QRSII departs from earlier versions by requiring that a network be drawn. The network tells QRSII about the difficulty of traveling between locations in an urban area. The urban area is described by a set of zones. Zones may vary greatly in sizes, but they can cover the whole area without leaving any gaps.

QRSII accomplishes its forecast by first determining the number of person-trips that begin and end in each zone. This step in the model is called "trip generation". The second step In the forecast is to determine for each trip purpose ( Home-Based-Work (HBW), Home- Based-Not-Work (HBNW), Not-Home-Based (NHB)) the number of person-trips that go from any given "origin" zone to any "destination" zone. This step in QRSII is called "trip distribution". The third and final step converts person-trips to vehicle-trips and assign them to the links in the network following the shortest paths. This step is called "traffic assignment".

The latest Windows versions of QRSII (7.0) and General Network Editor(7.0) was purchased. The new version greatly increases GNE's power, both as graphics tool and data preparation tool with these limitation:

2. INPUT DATA

The QRSII required, first of all, drawing the transportation network, then identifying each node and link of the network. A node could be identified as centroid, intersection with or without penalties, or production attraction tag. A link, however, could be identified as one way street, two street, centroid connector, or piece of string. Once the node and link identification process was completed, numerical data were entered.

The HIATS contains 397 traffic zones ( 88 zones in Ohio; 179 zones are in West Virginia; and 130 zones are Kentucky), 3000 nodes and 6000 links. 398 thru 425 are the external cordon line stations (28 external stations). Figure 1.2 shows the existing study area and highway network entered into the computer. Figure 1.3 shows the locations of the centroids and external stations, and the way they were connected to the highway network by centroid connectors. The original highway net work entered into the computer was composed of 2069 nodes and 2697 links . Out of the 2069 nodes, 398 nodes were identified as centroids, 28 as external stations, 429 production attraction tags which contained the trip generation (production and attraction) values of the centroids and external stations, and the remaining as intersections without penalties. Later on during the validation process , some of the intersections without penalties were changed to intersection with penalties. Apart from the centroid connectors and the pieces of strings which connected each production attraction tag to either a centroid or external station, the remaining links were identified as two way and one way streets.

Having completed the drawing and identification process, numerical data were entered. The base year of the continuing transportation study for which KYOVA is responsible is 1990. Thus, the 1990 base year trip generation values were used to validate the highway transportation network. Speed on the centroid connectors was assumed to be 12 mph. Speed and capacity are estimated to be 50 mph and 1000 vehicles per hour per lane (vphpl) respectively on freeways; 28 mph and 800 vphpl on the arterials; 18 mph and 500 vphpl on collectors. The population value must be entered so that QRSII can select the correct set parameters to be used for trip generation, trip distribution, and vehicle assignment. The period of time for forecasting trip making was selected to cover a whole day (24 hours).

QRSII does not have the capability to take into account the trips between external stations. To over come this obstacle, results of previous KYOVA studies related to through traffic were used.

2. OUTPUT DATA

The main result file of QRSII is called "output.dta". It is nearly identical to the input file, except volumes and intrazonal trips have been calculated and their results have been entered automatically into the result file. Also, QRSII provides many tables (see Appendix A: HIATS Network Files and File Description). The ones of interest for this research are mentioned below:

1- "NODELABL.TXT": this file contained the names of centroids and external stations.

2-"LINKLABL.TXT": this file contained the link identification number and the link label.

3-"LENGTHDS.TXT": this file contained the number of vehicle-trips for each purpose and the trip length distributions.

4. "LINKVOLS.TXT" : this file contained the volumes that have been assigned to links.

5." VEHTRIPS.TXT" : this file contained the vehicle trips that come from and go to each zone.

6."PSANDAS.TXT": this file contained the file contained the trip productions and attractions for each centroid and external station.

3. THE CALIBRATION PROCESS

QRSII separately distribution for each of its purposes: home-based work (HBW); home-based nonwork (HMNW); and nonhome-based (NHB). There is one parameter for each purpose. The value of this parameter is inversely related to the average length

Large parameters implied that origins tend to be close to destinations. Conversely, small parameters implied that origins tend to be distance from destinations.

In order to calibrate QRSII, the manual has suggested to use reciprocal of the average trip length as starting value of the parameter. After a few iterations, the calculated average trip length became reasonably close to those of the 2000 given earlier in this paragraph.

3. THE VALIDATION PROCESS

Having completed the input of all required data and finished the calibration process, QRSII was run. IT took about 45 minutes of execution time to complete the run to provide the 1990 free traffic assignment. The traffic volumes on the network provided by this 2000 traffic assignment were compared to those of the 1990 ground count traffic volumes. Penalties were imposed on the intersections to which the connected links had assigned traffic volumes different from those of the 2000 ground counts. It took few assignment runs to get an assignment that compared favorably to the 2000 ground count volumes. This process is called the validation process. Once the validation is completed, the 2030 forecasted trip generation (production and attraction) values replaced those of the 2000. At this stage, QRSII was run again to provide the "2020 do nothing" traffic assignment.

CONCLUSION

The rest of this report documents the various steps required to check (validate) the sequential travel demand models. Section 2 describes the required input data and their sources. Also, it provides a complete account of this assumptions made to develop the required input data. Section 3 deals with checking the validity of the trip generation phase of the sequential travel demand modeling process. Section 4 discusses the trip distribution phase, and Section 5 the assignment phase of the sequential travel demand modeling. Section 6 provides a summary of conclusions and recommendations and points out areas of concerns with regard to the 2000model validation efforts