Section 4: Policy Options Evaluation Tool
Every HOV lane is unique in its demand composition and operational characteristics. These characteristics are often difficult to quantify, so it is challenging for HOV operators to know exactly how well their HOV lanes are operating. Likewise, the impacts of any policy changes to their HOV facilities are also difficult to quantify, and would create additional uncertainty concerning future HOV performance. So before making any changes, it is critical to understand: (1) the current operating conditions of the existing HOV facility; (2) what impacts on the operational performance of the HOV facility can be expected with policy shifts; and (3) whether policy shifts will help the operator meet the goals and objectives established in the study region.
Travel demand modeling is one approach commonly used to evaluate current and future conditions in transportation systems. These models can be used to estimate the potential impacts of policy shifts, including changes in HOV lane policies. However, the traditional modeling process tends to be complex and requires extensive time and budget to implement, rendering it ineffective for quick-response analysis.
The Policy Options Evaluation Tool for Managed Lanes (POET-ML) was developed as one feasible alternative to travel demand modeling. The tool makes it possible for HOV operators to complete a current HOV system condition assessment, quantify the impacts of HOV lane policy shifts on operational performance and financial feasibility, and ultimately prioritize the most appropriate HOV policy changes, or combination of HOV policy changes, to best align with their system goals and performance objectives. This will all be accomplished through a simple user interface that does not require extensive modeling know-how. Users equipped with even limited input data will be able to apply what they know to get sketch-level planning output and suggestions for HOV policy modification.
Specifically, POET-ML has been structured to help HOV operators answer the following questions:
The POET-ML framework, methodology, and illustrative results from four typical scenarios (Scenario 1: HOV & GP Lanes Both Under Capacity; Scenario 2: HOV Lane Under Capacity & GP Lanes Congested; and Scenario 3: HOV Lane & GP Lanes Over Capacity [Increased Restrictions]; and Scenario 4: HOV Lane & GP Lanes Over Capacity [Additional Capacity]) are provided next. The POET-ML tool itself is provided as a separate deliverable.
Framework, Methodology, and Illustrative Results
Figure 4-1 illustrates the analytical process used in POET-ML.
Step 1: Operational Assessment of Existing HOV Facility
The initial step in the model process is an assessment of the operational effectiveness of the existing HOV facility. This assessment considers both physical and operational characteristics including number of lanes, length, separation, eligibility, and demand, among others.
In this step, the user can select a specific HOV facility from the FHWA Highway HOV Facility Inventory database that includes information on HOV policy details and physical characteristics. The user is then required to enter the number of HOV lanes and GP lanes in each direction during peak hour operations as well as the corresponding volumes in these lanes (records highlighted in red). Other information, such as public transportation vehicles (no. of buses per hour); percentage of motorcycles; percentage of taxi and percentage of low emission and/or energy efficient vehicles, is optional. Once valid values are entered for these items, the user can continue with the analysis. It is also possible to store a specific profile for future use by modifying the text for one or more of the input data field records.
Table 4-1 outlines the set of information to be populated either from the FHWA Highway HOV Facility Inventory database or by the user. The data was grouped into four major categories. General information, physical characteristics, and HOV policies should be readily available to nearly any user familiar with the HOV system under consideration. However, travel demand and operational performance could be more difficult to obtain.
* Volumes represent demand for the corridor by lane type.
User input, to be input to the model using the interface shown in Figure 4-2, will supply the information necessary to assign the HOV facility to one of two categories based on the established performance thresholds, such as volume-to-capacity ratios or service flow rate (pc/h/ln). These categories describe the general performance of the facility in terms of utilization. During step 2 of this process, the user will be presented with a set of policy adjustments based on the specific category to which the facility is assigned. Table 4-2 outlines the two categories and corresponding performance thresholds by default.
It is important to note that the default threshold values of V/C ratio (0.75) and service flow rate (1650 pc/h/ln) were established based on aggregated national survey results, and they are consistent with the assumptions in FHWA’s Spreadsheet Model for Induced Travel Estimation - Managed Lanes (SMITE-ML). The default values of V/C ratio and service flow rate are stored in the POET-ML parameters page and remain interactive and transparent to the user. Users are allowed to adjust these values to reflect the unique characteristics of facilities in their region. To review and/or modify the default model parameters navigate to the Potential Impacts page and select “Adjust Parameters.”
The precision of the analysis will depend on the availability of data from the user, and the quality of the final model output depends entirely on the user’s ability to supply as much needed information as possible.
The example corridor is I-85 in Atlanta, GA from I-75 north to SR316. This 24 mile facility has a single HOV lane in each direction with HOV2+ occupancy policy. Key information was loaded from the HOV data base. User input included:
Step 2: Identification of the required and/or optional HOV policy changes
A set of applicable policy adjustments are introduced in step 2 of the model process, based on the assessment from step 1. If it is determined that the HOV facility has excess capacity in both the peak and off-peak periods, the user will be shown a number of policy change options related to vehicle occupancy, vehicle eligibility, and pricing. In order to fill unused HOV capacity, and avoid empty lane syndrome, the user could choose to lower the occupancy restrictions (e.g. from HOV3+ to HOV2+) or to allow additional free vehicles (e.g. public transportation vehicles, taxis, motorcycles, hybrid vehicles, etc.). Additionally, the user could convert the lanes from HOV to high-occupancy toll (HOT) lanes, and sell excess capacity to users not permitted in the lanes but who would be willing to pay for the travel time savings these lanes provide. These policy changes could also be bundled together in some combination that both achieves the utilization targets and meets the goals of the region. Table 4-3 shows the options to be presented to the user.
The same set of policy change options applies for HOV facilities determined to be congested during peak periods. However, the potential adjustments will be more restrictive, rather than less restrictive, as was the case for the excess-capacity scenario. For example, one option to address congested HOV lanes is to increase the occupancy requirements (e.g. from HOV2+ to HOV3+). Likewise, non-carpools that are currently eligible to use the HOV lanes could be prohibited (e.g. disallow motorcycles, transit vehicles, etc.). Pricing of non-eligible vehicles can also be implemented on congested HOV lanes, but it must be bundled with some other policy shift. Once demand in these lanes is brought down below capacity through more restrictive policies, any remaining capacity could be sold to ineligible vehicles (i.e. those not meeting the current occupancy/eligibility policy) that are willing to pay for access. In addition to policy change options related to vehicle occupancy, vehicle eligibility, and pricing, the user can also explore the impacts of adding an additional managed lane. This option is only available for HOV facilities that are congested during peak period. This could either be an additional lane in each direction, or an additional reversible lane, depending on the facility. By adding additional capacity, it provides increased flexibility for HOV operators and eliminates the need for immediate occupancy policy changes. Table 4-4 shows the options for the congested peak period condition.
If HOV demand is deemed to be on target during peak periods (i.e. neither underutilized nor congested), there are still opportunities for policy adjustment. Future demand may eventually lead to congestion in lanes that are operating well today, and proactive steps could ensure efficient operation for years to come. Pricing is always an option that provides flexibility for HOV operators to manage demand in these lanes in order to achieve more efficient use. Occupancy and eligibility policy changes alone, offer only discrete solutions that may tip the utilization balance too far in one direction.
Based on the volumes in the corridor, both the HOV and general purpose lanes operate at undesirable levels, LOS E and F for the HOV and GP lanes respectively.
Potential policy adjustments include:
Step 3: Evaluation of Potential Impacts
The third step in the process is to assess the impacts of the HOV lane policy change or combination of policy changes that were selected in step 2. The tool will track four key measures of effectiveness: travel demand impacts, mobility impacts, environmental impacts, and financial feasibility.
Travel Demand Impacts
Both vehicle and person travel demand will be examined over daily and peak hour periods in the HOV/HOT and general-purpose (GP) lanes. Travel will be broken down into carpools, transit, motorcycles, special fuel vehicles, taxis, and paying vehicles. At a minimum, the user will be required to supply information on peak hour vehicle trips for each vehicle type under the current HOV policies. Relationships coded into the tool will be used to calculate peak hour person trips and daily vehicle and person trips.
Travel demand impact calculations will depend heavily on which of the two conditions (excess capacity or congested) applies to the facility under evaluation. If pricing is selected as a policy change, the level of travel demand in priced lanes will be maintained at Level of Service C during the peak hour, by default, i.e., about 75% of absolute capacity. Paying vehicle volumes in priced lanes during the peak hour are estimated to be equal to the spare vehicle capacity that would be available on the lanes at a Level of Service C. The balance of the volume is occupied by non-paying vehicles.
A number of combinations exist between existing conditions and subsequent policy adjustments. The algorithms in place to determine final volumes for both HOV/HOT lanes and general purpose (GP) lanes are different based on the combination under consideration. Following are four potential scenarios, meant to outline the different calculation processes executed by POET-ML. Each scenario description includes a table with example output data and a figure showing general travel conditions in the corridor. Following these scenarios is a detailed description of the calculations for mobility and environmental impacts, along with financial feasibility.
Scenario 1: HOV & GP Lanes Both Under Capacity
Many corridors with HOV lanes are uncongested in peak periods. Under these conditions, no changes are required to bring HOV operating speeds back to acceptable levels. However, the HOV operator may be interested in seeing the impact of implementing pricing, or of allowing additional vehicles into the HOV lanes through occupancy or eligibility changes. Figure 4-3 illustrates the potential impact of allowing priced vehicles into the HOV lanes. The colored arrows represent the flow conditions for each lane in the corridor. Table 4-5 shows an example calculation for this scenario.
Of the 1,100 peak hour HOV trips in the existing condition, 1,004 of them are carpools. The rest are other eligible vehicles. These other vehicles generally make up a small proportion of total HOV demand, and therefore changes to eligibility restrictions could have little direct impact on HOV and GP lane performance.
Initially, this uncongested corridor experiences LOS C conditions in the GP lanes and LOS A/B conditions in the HOV lane. Allowing priced vehicles in the HOV lane will attract additional users because of the time savings relative to the GP lanes. POET-ML pulls these users from two different places: the GP lanes and parallel facilities. The percent split from these sources depends on the conditions in the GP lanes. As the V/C ratio in the GP lanes rises, the contribution of vehicles from these lanes to the HOV/HOT lane also rises. When GP conditions are near LOS A/B, a larger portion of vehicles are diverted from parallel facilities to the HOV/HOT lane. The final volumes in the HOT lane under this condition are no higher than the maximum LOS C volume defined in the Tool. Nor are they larger than ¼ the total corridor volume (¼ because the facility has 4 total lanes, with one HOV lane). This is to ensure that HOT operating speeds do not fall below GP speeds, which is a possible, but unlikely scenario. For these reasons, revenue is likely to be minimal under this condition. Obviously, few motorists would be willing to pay a toll to use the HOT lane when only minimal time savings can be realized.
Indeed, Table 4-5 shows just 300 paying vehicles after the policy change, bringing the peak hour total in the HOV/HOT lane to 1,400. Volumes decrease from a total of 4,500 on the GP lanes to 4,380. With a per lane capacity of 2,200 vehicles per hour, the GP lanes have a similar V/C ratio to that of the HOV/HOT lane, which is the reason for the low demand from paying vehicles in that lane.
Scenario 2: HOV Lane Under Capacity and GP Lanes Congested
Another common scenario involves congested GP lanes adjacent to an HOV facility that operates well below capacity. Again, the operator is not required to make policy changes in order to maintain an acceptable LOS in the HOV lane, but there may be interest in achieving greater utilization in this lane. Options for increasing HOV volumes include relaxing occupancy and eligibility restrictions in the lanes, as well as allowing previously ineligible vehicles (e.g. single-occupant vehicles) to pay a toll in order to use the lane. These options would have different impacts on lane volume, however, and caution needs to be observed to avoid creating congested HOV conditions. For example, lowering the occupancy restriction from 3+ to 2+, if applicable, could potentially allow too many vehicles into the HOV lane, degrading performance below acceptable levels.
In this example, a congested corridor has an underutilized HOV lane. This condition is commonly referred to as “empty lane syndrome,” and is one key motivator for HOV policy change. Allowing priced vehicles access to the HOV lane can lead to improvements in the GP lanes and better use of the HOV lane. One likely outcome of this change can be seen in Figure 4-4. Here, LOS improves from ‘E/F’ to ‘D’ on the GP lanes, while LOS degrades slightly on the HOV/HOT lane from ‘A/B’ to ’C’. In POET-ML, most of the priced vehicles in the HOT lanes come from the GP lanes under these conditions, with a small contribution from parallel facilities. As a result, total corridor throughput increases slightly under this scenario. As noted previously, vehicle contribution from these two sources is determined based on a sliding scale with a 70/30 split between parallel facilities and GP lanes when the GP lanes operate at LOS A. This split changes to 60/40 under LOS B, 50/50 under LOS C, 40/60 under LOS D, and 30/70 under LOS E/F conditions. This distribution is included in the parameters page, and can be modified by the user.
HOT lane volumes are capped at the LOS C capacity, which is accomplished in practice through demand-responsive, variable tolling. If pricing is not a viable alternative, an HOV operator could still achieve greater corridor throughput by increasing the types of eligible vehicles in the HOV lane. Allowing hybrids or special fuel vehicles, taxis, or additional transit vehicles can provide a degree of relief to the GP lanes while increasing utilization of the HOV lane. However, as discussed previously, relaxing eligibility restrictions may not impact many vehicles, and therefore conditions may not change much in the corridor.
Scenario 3: HOV Lane and GP Lanes Over Capacity (Increased Restrictions)
Some HOV facilities are congested during peak periods and require policy adjustment in order to maintain federally mandated performance standards. Low cost strategies for decreasing HOV lane volume include increasing occupancy restrictions and implementing more exclusive eligibility criteria. However, efforts to divert vehicles from the HOV lanes can lead to increased congestion on GP lanes. And if HOV lane rules are made too restrictive, traffic could fall well below LOS C conditions, leading to empty lane syndrome. For example, in many urban areas, the vast majority of HOVs have just 2 occupants, with only a small percentage of 3+ occupant vehicles. If the HOV operator increases the occupancy restriction from 2+ to 3+, many of the vehicles in the lane will no longer be eligible, and will be diverted to the GP lanes or parallel facilities.
POET-ML takes into account the overcapacity scenario described above. Once it is determined that the HOV lane is congested in peak periods, the user is presented with a list of potential policy changes designed to achieve improved HOV operating conditions. The greatest impact usually comes from increased occupancy restrictions. Figure 4-5 shows an example of the impact of first increasing this restriction from 2+ to 3+, followed by allowing priced vehicles in the lane.
The first selection shifts a large number of vehicles from the HOV lane to the GP lanes. Of course, the number of vehicles diverted will vary by facility, according to the regionally-specific HOV mix (i.e. the relative number of HOV2, HOV3, HOV 4+, etc.). This split is coded as a parameter in the Tool, and it can be updated by the user as desired. If the user changes only the occupancy restriction, total corridor volume will remain constant, and GP lane conditions will likely become even more congested. In addition, it is possible that the HOV lane may exhibit LOS A/B conditions, which is suboptimal utilization. If the user follows this selection by allowing pricing in the HOV lane, however, vehicles return to the lane and fill the unused capacity. POET-ML pulls most of the priced vehicles from the GP lanes, and a smaller portion from parallel facilities. This split is also coded as a parameter in the Tool, and if the user desires to vary the source of priced vehicles, he or she has that flexibility. Once both decisions are executed, conditions are likely to appear as they do on the right of Figure 4-5.
Table 4-7 shows the extent to which the GP lanes become more congested in this scenario. Of course, the HOV lane is maintained at the LOS C capacity, and most of these vehicles are tolled. HOV3+ vehicles, along with other eligible free vehicles, comprise the balance of the lane volume. The HOV2 vehicles, which were pushed to the GP lanes in response to the occupancy policy change, are responsible for the increased GP lane congestion.
Scenario 4: HOV Lane & GP Lanes Over Capacity (Additional Capacity)
Another option for addressing a corridor with congested HOV and GP lanes is to add HOV/HOT capacity. In locations where available right of way affords such an investment, this option can provide a high degree of flexibility for HOV operators. Additional HOV capacity allows greater opportunities for efficient corridor flow and can eliminate the need for occupancy and/or eligibility policy change. In the scenario highlighted in Figure 4-5, the user has opted to add HOV capacity and implement tolling in these lanes. In doing so, corridor conditions are improved for both the managed and GP lanes. In addition, total corridor volume increased, which means the facility can serve more vehicles, more efficiently than before. And all of this is possible while maintaining occupancy restrictions of 2+. This last point is important, because raising occupancy restrictions can be controversial. Those that have formed 2 person carpools to use the HOV lanes will likely object to any change in policy that forces them out of the lanes. Additional capacity can help avoid this situation.
In the scenario highlighted in Figure 4-6, the user has chosen to address corridor congestion by maintaining the existing HOV policy, adding a lane of HOV/HOT capacity, and implementing pricing on both lanes. POET-ML is equipped to respond to each of these decisions and to calculate the final conditions on the managed and GP lanes. The additional HOV lane doubles the capacity for qualifying vehicles. These vehicles are spread evenly over the two lanes, which likely eliminates peak period congestion. Allowing priced vehicles fills unused capacity in these lanes while helping to improve conditions in the GP lanes. Again, the majority of paying vehicles are taken from the GP lanes, with a smaller percentage diverted from parallel facilities.
Additionally, POET-ML analyzes the peak hour person trips based on occupancy rate for different vehicle types. It also analyzes the total daily vehicle trips and total daily person trips based on Peak Hour vehicle/person trips and daily to Peak Hour Conversion factor. Table 4-9 outlines those travel demand impacts and its corresponding calculation methodology. Table 4-10 fills in these formulas, under the existing conditions, with values from the example cited earlier.
The travel demand impacts will then be used to determine the facility operating conditions, including the volume-to-capacity ratio, operating speed, level of service, facility travel time, total vehicle travel delay, etc. Again, these impacts will be examined over daily and peak hour periods for both the HOV and GP lanes.
To calculate each of the mobility impacts, a number of assumptions are embedded into the calculations of these impacts. Examples of values the user may wish to update include Hourly Freeway Capacity per Lane (vph), Free Flow Speed (mph), the values for “alpha” and “beta" used in the Bureau of Public Roads equation for computing congested Peak Hour and Daily Travel Speeds, V/C thresholds for Level of Service, etc. All these assumptions are stored in the POET-ML Parameters Page. The user will have explicit access to change these assumptions if desired to better fit the characteristics of specific facilities and areas.
The calculations for mobility impacts are built with flexibility in mind, allowing the user to customize the assumptions to a specific region, if the data supports it, and if there is a desire for greater precision in the results. If not, the user can work with the set of assumptions that emerged out of the model calibration effort, which will be based on nationwide averages.
Table 4-11 outlines the information to be included for these mobility impacts and the detailed calculation methodology of those mobility impacts. Table 4-12 fills in these formulas, under the existing conditions, with values from the example cited earlier.
POET-ML will use the traffic volume estimates and mobility impact estimates to evaluate the environmental effects of the HOV facility under consideration. Two key environmental indicators will be examined, including air quality performance and carbon dioxide.
The quantity of gasoline conserved can directly relate to reduced vehicular emissions. Gasoline savings were based on numbers derived using Texas Transportation Institute assumptions of 0.68 gallons of fuel per hour of delay. The evaluation will consider changes in total vehicle delay as a result of the policy adjustment package selected by the user and estimate the fuel-based emissions based on gas savings and estimated vehicular emission rates per gallon. Due to the difficulty in determining advancement in emissions technology, the values used in POET-ML reflect modern day estimated emission rates, as illustrated in Table 4-13.
For example, if the model results predict a reduction in total vehicle travel delay as a result of a conversion form HOV lanes to HOT lanes, POET-ML will model changes in air quality and carbon dioxide emissions based the above average rates. The user will be able to see the impact of any delay reduction in environmental terms.
Table 4-14 outlines the information to be included for the environmental impacts and the calculation methodology of these two performance measures. Table 4-15 fills in these formulas, under the existing conditions, with values from the example cited earlier.
Again, the user will have access to adjust all the estimated emission rates for CO (kg/gallon), NOx (kg/gallon), VOC (kg/gallon) and Carbon Dioxide (kg/gallon) in the POET-ML Parameter page if desired.
The final measure of effectiveness is financial feasibility. The set of policy adjustments includes pricing of existing HOV lanes, and if pricing is selected, it will trigger additional analysis in the model. Again, the user will have access to assumptions behind these calculations, including value of time estimates, weekend/weekday revenue ratios, and per-transaction tolling operation costs. The model incorporates national averages for these inputs and uses the results from the mobility impact analysis to perform this financial evaluation.
Output for this step includes the number of tolled vehicles, daily and annual revenue, and annual toll operation costs. Also bundled with this output is a set of transportation benefits calculated in monetary terms. These are presented as travel time and fuel savings as well as daily user benefits resulting from policy changes implemented on the HOV facility. Specific measures of financial feasibility and its corresponding calculation methodology are listed in Table 4-16. Table 4-17 presents example calculations assuming that occupancy restrictions were changed from 2+ to 3+ in the example cited previously.
United States Department of Transportation - Federal Highway Administration