Continuous Dynamic Optimization: Impacts on ADA Paratransit Services (2023)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

1   Why This Synthesis Was Commissioned Continuous dynamic optimization (CDO), as applied to Americans with Disabilities Act (ADA) paratransit services (and other demand-responsive services), is an automated process by which a scheduling/dispatching technology continuously or frequently considers addi- tional trips just booked, changes to booked trips, cancellations, and day-of-service events in solving problems or in taking advantage of opportunities. In view of those changes, the CDO process then re-optimizes the assignment of trips to a transit agency’s dedicated vehicle runs and to the available vehicles of overflow non-dedicated service providers to achieve the transit agency’s desired balance of service/cost efficiency and service quality. As the industry has witnessed, the initial and ongoing involvement of a transit agency’s technology partner has proved to be as important as the CDO technology itself. The development of CDO as part of paratransit scheduling and dispatching systems origi- nally stemmed from a desire to more effectively schedule trips that are left unassigned after the schedulers have left for the day and to address reductions in productivity and on-time perfor- mance that occur on the day of service. For example, events that happen on the evening before the day of service and throughout the day of service have a domino effect on productivity and on-time performance. These common events can include the following: • Advance cancellations [typically defined as cancellations received two hours before the scheduled pickup time (or window) after the scheduling process has concluded]. • Late cancellations and no-shows during the service day. • Drivers who call out on the morning of service and cannot be replaced, requiring the reassignment of trips from the run of those drivers to one or more other runs. • Vehicles downed at the pre-trip inspection that cannot be replaced, requiring the reassign- ment of trips scheduled that run to one or more other runs. • Reemerging no-show trips and will-call return trips. • In-service vehicle breakdowns and incidents. • Drivers running behind schedule. Most ADA paratransit systems have a dispatching staff that addresses the changes that need to be made because of these events. The focus of most of this staff, however, is solving current problems, such as the need to reassign trips because of a driver callout or to reassign an upcoming trip from a run that is behind schedule to get that run back on schedule. Unassigned trips that fall out of the scheduling process are typically scheduled by the time the last scheduler goes home for the evening. Many times, however, the early morning dis- patcher will have to schedule a few unassigned trips left over from the night before because of unavailable slots. S U M M A R Y Continuous Dynamic Optimization: Impacts on ADA Paratransit Services

2 Continuous Dynamic Optimization: Impacts on ADA Paratransit Services The dispatching staff of some ADA paratransit services (most often the lead dispatcher) also performs proactive dispatching. This process involves looking ahead in the schedule at potential problems that stem from the previously mentioned events that affect productivity and on-time performance to address/solve the problems before they manifest themselves in real time. With some software systems, these types of problems are red flagged to alert the dispatcher; a skilled proactive dispatcher knows which ones need to be addressed and which ones will likely resolve themselves. In addition, the proactive dispatcher is often on the lookout for opportunities to improve productivity or to fill holes left by cancellations. Proactive dis- patching requires a particular focus and skill set and, if done well, can be hugely important to ADA paratransit services because it is one of the few things that have a positive impact on both productivity and on-time performance, which normally have an inverse relationship. Unfortunately, many transit agencies that offer ADA paratransit service cannot perform proactive dispatching because they have just enough staff to deal with the real-time problems that need the immediate attention of a dispatcher. And even with the best intentions, a pro- active dispatcher may get dragged into these problems, losing focus on potential problems. This challenge was the primary impetus for paratransit software vendors to develop CDO capabilities—to automate the proactive dispatching function by continuously optimizing the current day’s schedule in consideration of changes that occur after the scheduling process has been completed and through the end of the service day. Such changes can include rider cancellations and no-shows, late-running vehicles caused by unforeseen traffic congestion, an in-service breakdown, and a run that is removed from service (before it starts) as a result of a driver callout or a safety issue caught in the pre-trip inspection. These CDO systems automatically consider changes such as these in scheduling unassigned trips, solving upcoming problems ahead of time while continuously looking for new opportunities to improve produc- tivity, on-time performance, or both. The software systems with CDO generally have configurable intervals regulating how often this process is executed; depending on the speed of the process, the interval could be set so that optimization is almost continuous. These systems generally also provide configurable safeguards against modifying a rider’s pickup time or interrupting what a driver is doing at a given moment by specifying a configurable offset of hours, which means that the systems do not modify the manifest of an active/in-service run for the next hour or two. Motivation for developing CDO capabilities was also tied to the labor-intensive and chal- lenging function of paratransit dispatching and especially proactive dispatching. By pre- empting upcoming problems before they manifest and taking advantage of opportunities that will allow the system to further optimize the schedule, CDO preempts and thereby reduces the number of real-time problems that dispatchers have to manage. In this way, CDO frees up dispatcher labor to address more immediate tasks, such as processing no-show requests and responding to immediate driver and customer needs. Moreover, because CDO is automated, it offers several other advantages. First, it is inher- ently more objective than perceived “dispatcher favoritism,” which often causes friction between dispatchers and (other) drivers. Second, the automated aspect of CDO could reduce the number of schedulers needed if a transit agency is confident that CDO will take care of the trips left unassigned after the scheduling process has concluded. Third, relaxing sched- uling parameters in subsequent batch-scheduling processes in order to accommodate more trips—a common scheduling tactic—may no longer be necessary if the CDO capability cor- rects for this. Fourth, CDO means that service will be of a higher quality if compromises used in the scheduling process can be minimized or eliminated. Finally, some agencies have replaced the traditional scheduling function associated with next-day and advance reserva- tion trips with CDO.

Summary 3   The point of this synthesis and in particular the case examples is to document these advantages and discuss how transit agencies confront and address challenges so that other transit agencies can benefit from their lessons learned. Some of the more traditional paratransit scheduling software systems, such as Ecolane, Trapeze, and QRyde, are used by transit agencies to deploy the software’s CDO in support of their ADA paratransit services. In addition, other vendors, such as Via Mobility, Spare Labs, and RideCo, provide the transit industry with on-demand technology (if not turnkey and brokerage contracts) primarily for microtransit. These vendors now have software that supports paratransit services, and their transit agency users employ this software, with its CDO capabilities, in support of their ADA paratransit services. Another vendor, Moovit, is following suit. The importance of transit agencies now using not only on-demand system technologies for ADA paratransit and their microtransit service(s) but also ADA-paratransit- certified drivers for both is that these resources, when consolidated, can result in more effi- cient service. Summary of the Synthesis Effort Because there is very little literature on the impact of CDO on ADA paratransit services, the Texas A&M Transportation Institute (TTI) conducted an industry scan to learn from sources across the country about the different approaches to providing these services and how these approaches are informed by different service models and circumstances. Much of this information came from the marketing materials of the vendors themselves. Through the vendors’ and TTI’s knowledge of the industry, TTI identified 24 transit agen- cies in which CDO was reportedly being used to support ADA paratransit. Meanwhile, the SB-36 survey questionnaire was developed and then revised with feedback from the project panel. An online survey program called Qualtrics was used for the survey; in this way, TTI needed to provide only the survey link to transit agencies willing to participate. Telephone calls were made to contacts at each of these transit agencies to verify that they were indeed using CDO for their ADA paratransit services and to determine their willingness to participate in the survey. In addition, two of the technology vendors, in order to guard the privacy of their users, sent them the link to the survey on behalf of TTI. However, the response to these vendor requests was poor. Moreover, some of the transit agencies that did respond to the survey were not using CDO to support ADA paratransit services. Ultimately, 24 transit agencies responded to the survey. However, some of them indicated that they were not using CDO in their daily activities or were using CDO but not in support of an ADA paratransit service; other agencies started but did not complete the survey. While this left only 11 transit agencies, their ADA paratransit services collectively used four technologies (provided by Ecolane, Trapeze, Spare Labs, and Via Mobility) and had different experiences with CDO. The 11 transit agencies also reflect diversity in location, size, and service model. Chapter 2 describes the findings from the survey. Because of the limited response to the survey, a decision was made to revise the work plan: instead of developing five full case examples, TTI created service profiles (case examples) of all 11 agencies that completed the survey. In preparing these profiles, TTI interviewed contacts at the 11 transit agencies to clarify and expand on their survey responses. Chapter 3 describes these profiles. As part of this study, TTI also interviewed staff from several technology vendors whose CDO capabilities support ADA paratransit services. These vendors are Ecolane, HBSS/QRyde, RideCo, Spare Labs, Trapeze, and Via Mobility. Some of the interviews led to an online

4 Continuous Dynamic Optimization: Impacts on ADA Paratransit Services demonstration of the technology. TTI staff also attended a Trapeze users’ conference, at the vendor’s invitation, to interview staff and users and to participate in hands-on training. Finally, TTI put together a synthesis of lessons learned from the survey, the profiles, and the vendor interviews. Chapter 4 provides this information. Summary of the Findings Benefits of CDO Technology Virtually all transit agencies involved in both the survey and the profiles were enthusiastic about the use of CDO as a way to automate proactive dispatching and, in some cases, to auto- mate the scheduling function. Common motivations for using CDO include the following: • To improve productivity. Eight of the 11 transit agencies in the study specifically acquired technology that had CDO capability to improve productivity, as measured by passenger trips per revenue vehicle hour. In specific, the transit agencies acquired the CDO technology to counteract the typical decreases in productivity that arise from driver callouts, downed vehicles, late cancellations, no-shows, incidents, breakdowns, and vehicles running late because of unforeseen traffic congestion and other factors. Six of the 11 agencies reported that CDO worked, with an average improvement of 17 percent in productivity; these six included one transit agency that achieved an improvement in productivity of 31 percent. • To reduce overall costs. Six of the transit agencies indicated that the change in operating costs ranged from a 30 percent decrease to a 4 percent increase, with the average change in operating costs amounting to a 13 percent reduction. The reduction in operating costs was largely a result of a decrease in the number of vehicles, in the revenue vehicles’ needed hours, or both, while the reduction in unit operating costs stemmed from being able to serve more trips with fewer resources. • To improve on-time performance. Seven of the 11 transit agencies indicated that improving on-time performance was another motivator in acquiring CDO. Six of these agencies indicated that their on-time performance did, in fact, improve, with one transit agency indicating a 9 percent improvement. And for those agencies that experienced a decrease in on-time performance (with an average decrease of 1 percent), the underlying cause was a significant increase in productivity. • To address dispatching problems. Eight of the 11 transit agencies saw CDO as a way to improve the quality of the dispatching. Four of the 11 agencies indicated that their dispatch staff was previously overburdened, and one agency indicated that its proactive dispatcher did not have the requisite skill set for this task. Moreover, eight of the transit agencies (73 percent) did report a reduction in real-time problems. In addition, a few of the agencies indicated that the objectivity of the optimization addressed some of the dispatcher-driver favoritism issues that often plague paratransit services. While not being primary motivators for purchasing the technology (as indicated by the survey responses), the other benefits that have accrued from CDO include the following: • A reduction in or repurposing of scheduling/dispatch staff and a reduction in scheduler/ dispatcher labor costs. • A reduction in customer complaints. Applicability of CDO The type of service model did not seem to have a significant impact on the applicability of CDO. For example: • CDO works for transit agencies that use it to support their ADA paratransit service. CDO also works for agencies that use it for coordinated/consolidated services that commingle

Summary 5   ADA paratransit with trips sponsored by human-service agencies (including Medicaid non-emergency medical transportation trips) and trips for the general public. • CDO is used both by transit agencies that directly operate the service and by transit agen- cies that retain turnkey contractors and operations contractors. CDO also works well with a split structure in which the transit agency handles the rider interface (bookings and estimated time-of-arrival calls), while contractors and overflow providers deliver services. • All of the CDO technologies used by the transit agencies are able to handle brokering to non-dedicated service providers that supplement the dedicated fleet or fleets. Impact of CDO on Scheduling Policies and Staffing One of the most interesting findings of the synthesis was that CDO enabled transit agen- cies to reconsider the way they do business. This included changing the way real-time scheduling is performed and changing advance reservation policies. Moreover, transit agencies began to see the possibility of replacing not only much of the service-day dis- patch effort with CDO but also the traditional scheduling process (and its separate staff) with CDO on the day (or days) leading up to the day of service. In many cases, this change has led to the repurposing of scheduling staff and a refocusing of the dispatch staff on real- time issues. Related Issues in Tailoring, Deploying, and Using CDO The experience of agencies in the study suggests that other transit agencies need to under- stand not only how relying on CDO may change the way in which the scheduling and dis- patching functions are carried out but also how the parameters can be shaped to provide the desired balance between service/cost efficiency and service quality. This balance can be achieved first by inputting hard numbers reflecting policy, such as the span of the pickup window, the default loading and unloading times for ambulatory versus non-ambulatory rides, and the maximum onboard time. Second, the optimal balance requires the tuning of parameters to reflect the realities of the service area and environment, and the transit agency’s goals. The tuning of scheduling parameters defines how important they are so that the system knows how to optimize the schedule. For example, how important is it that the scheduled pickup time is in the confirmed pickup window, or that there are no travel times longer than the maximum onboard time? Other parameters that are linked to service/cost performance include the minimization of total and deadhead miles traveled, the minimization of vehicle runs, the revenue service hours, or some combination of the three; and the maximization of shared rides. Another parameter is the extent to which a vehicle backtracks or zigzags to pick up or drop off another passenger before the system decides to put that other potentially ride-shareable trip on another run. With the tuning of these parameters, there is no right or wrong; there are only gradations. And the optimal setting can sometimes be achieved only by trial and error, which can be a time-consuming process of experimentation. Some of the CDO supporting systems allow agencies to directly access to the scheduling parameters; others do not. For the latter set, the vendor’s technicians are the only ones who can make adjustments (based on how they interpret the agency’s input, or feedback, or both). In addition to weighting the importance of the parameters, there are other ways to adjust the CDO processes. For example, almost all of the transit agencies indicated that their CDO processes involved being able to screen out or include certain trips in a particular optimiza- tion process, to anchor the pickup times for selected trips, and to order sub-processes such as optimizing wheelchair trips first. Many of the transit agencies in the study also indicated that their CDO processes involved the relaxation of some service quality standards.

6 Continuous Dynamic Optimization: Impacts on ADA Paratransit Services Transit agencies and vendors must consider the learning curve for both drivers and dis- patchers as the CDO technology is planned, implemented, and deployed. Some of the tech- nologies have simulation software for driver-training purposes through which drivers can get used to how the tablets work, how to perform events (time-stamping and geo-stamping arrives and departures at each stop), how to submit a no-show request, how to retrieve naviga- tion assistance, and so forth. For larger systems, the technology vendors suggested deploying the technology in groups of drivers rather than deploying it all at once systemwide. In this way, the parameters and the training can be adjusted incrementally based on initial feedback from the drivers. Lessons Learned Lessons learned from the survey and the profiles include the following: • Do not be afraid of CDO, especially because the benefits can be substantial. • Choose a technology partner that focuses on the transit agency’s needs not only during the transition to CDO but on an ongoing basis. • Define expectations and goals beforehand and prepare to adapt processes and practices. • Understand how the CDO capability works and how it controls drivers and vehicles. • Take the time for dispatchers and operations staff to learn to trust CDO. • Prepare for changes in staff composition and responsibilities. • Recognize that dispatchers will still be needed, but their tasks might be refocused. • Prepare for changes to operator shifts and attrition. • Turn off the CDO process if the dispatcher-to-driver communication system goes offline or if a tablet malfunctions. Need for Additional Research The size of the ADA paratransit services supported by the newer CDO technologies is increasing. As this trend continues, future research documenting the efficacy of these tech- nologies to support larger and more complex ADA paratransit systems will be needed. Many transit agencies are transitioning to these newer technologies to support their ADA paratransit services and their new or expanding microtransit services, all in one tech- nology platform. Interestingly, when these services are consolidated or if they overlap, ADA paratransit customers are choosing to use these microtransit services instead of the ADA paratransit services for some of their trips simply because microtransit services offer on-demand service. There is a need to research this dynamic as well as the extent to which one technology platform’s CDO process considers overlapping services in placing a trip. Report Organization After this summary, the synthesis report is organized as follows: • Chapter 1: Overview of the Technology. • Chapter 2: Survey. • Chapter 3: Transit Agency Profiles. • Chapter 4: Conclusions and Lessons Learned. • Appendix: Survey Instrument.