Text-Only Version Prepared by: TranSystems Corp. Medford, ma and: Planners Collaborative Boston, ma august 24, 2007 contents

• 
  Overall, scheduling within the system appeared to be good. Scheduling at Kiessling, which is mainly manual, appeared to be very thorough. Each of the other providers also performed a reasonable degree of manual review of the schedules automatically generated by the ADEPT system. A review of a sample of schedules also indicated that the sequencing of pickups and drop-offs was typically logical. If issues with circuitous routing are happening, it is likely that this is due more to same day adjustments and the dispatching process.
  
• 
  In some cases, particularly at JV and VTS, scheduling appeared to be very tight. Although “doable” on paper, little slack or recovery time was included in the schedules. Service providers appeared to rely on cancellations (which are considerable) to create slack time. Cancellations did not always occur, though, at the times and in the locations where slack was eventually needed. Tight scheduling with little slack/recovery places considerable pressure on drivers and the dispatch portion of the operation to handle same day service issues. It was recommended that the MBTA encourage service providers to build some slack time into schedules every few hours of the day. This should be done particularly on runs in outlying areas where opportunities to provide back-up in a timely way may not exist to the degree that it does in areas with a higher concentration of vehicles.
  
• 
  While scheduling at KTI appeared to be good, it is mostly manual. It was recommended that KTI work to make greater use of the ADEPT system and to utilize some of the features of that software. Final schedule adjustments and clean-up could then be manual. Over time, as demand grows, it is likely that it will become more difficult to manage the scheduling process is some level of automation is not used.
  
• 
  It was noted that JV did not have enough capacity during afternoon peak hours. Additional capacity needed to be added during the afternoon between 2:00 and 5:00 p.m. Afternoon capacity at VTS also appeared to be somewhat limited. As a result, trips are unscheduled and left on a “Wait List.” It also appeared to put pressure on schedulers to schedule tightly and use all available slack time on runs. This put pressure on dispatch to handle wait listed trips on the day of service – along with other significant dispatch duties – and to have to move tightly scheduled trips should any number of same day issues (traffic, longer than expected boarding times, etc.) occur. It also caused the system to rely heavily on cancellations to provide needed same day slack time. The reality, though, is that cancellations do not always occur at the times or in the geographic locations where slack is needed. It was recommended that JV and VTS explore ways to augment afternoon peak hour capacity to be able to schedule all requested trips.
  
• 
  It also was noted that service providers needed to be careful about scheduling pick-ups too close to the appointment time. Given a 30-minute pick-up window, pick-ups for short trips should not be scheduled too close to the appointment times, even if this means that riders may sometimes arrive 20-30 minutes early.
  
• 
  At the time of the on-site visits, schedulers appeared to have to do a lot of manual re-entry of trip information as they were clearing the wait list, and “cleaning-up” and fine-tuning schedules before call-backs are done. This was necessary because ADEPT locks-in promised times after the initial batch is run. To then assign new times to a trip in the manual scheduling process, the trip had to be re-booked. It seemed that this could result in data entry errors and even instances where trips are dropped due to manual scheduling errors. It was recommended that the MBTA work with StrataGen to see if there is a way to leave promised times flexible after the initial batch is run and then to lock all times when schedulers complete their manual clean-up and fine-tuning and before the call-back process is initiated.
  
• 
  It was recommended that the MBTA encourage all service providers to show pickup windows and appointment times on driver manifests. This is done at VTS. Having these times on the manifests would allow drivers to know the leeways they have in making pickups and in getting riders to appointments on-time.
  
• 
  It was recommended that the MBTA encourage service providers to make greater use of the ADEPT feature that allows for travel speeds to be refined and set for small areas known to be problems. This could be done partly by having Road Supervisors record actual travel times for areas known to be problematic. Several observations could be made over time and the information from these observations could then be used to refine travel speeds in the system. Until this is done, service providers need to be carefully checking the workability of trip times generated automatically.
  
• 
  It was noted that JV would benefit from additional scheduling staff on Sundays and Mondays to assist in developing Monday and Tuesday schedules. In 2005, only one scheduler was scheduled to work on these days.
  
• 
  It was recommended that service providers develop procedures that would allow them to plan for and adjust capacity to actual demand. This might be done by scanning the number of trip requests received three days in advance, two days in advance, and then mid-day on the day before service. Trips on file might even be “pre-batched” to get a sense of the capacity that may be needed. On-call drivers could then be maintained and called in if added capacity was likely to be needed. This would give schedulers greater ability to add runs when needed and supervisors time to arrange for additional drivers and runs. Alternatively, service providers could develop some flexible capacity (like VTS and GLSS) which could be called on when the actual demand cannot be met by the set run structure.
  
• 
  Finally, it was recommended that GLSS set its “scheduling window” parameters in the ADEPT system to be consistent with MBTA policy, which is to not schedule return pick-ups more than 30 minutes after the requested time, or drop-offs 30 minutes before the appointment time. At the time of the on-site visit, the GLSS parameters appear to be set to allow a 60 minute variance.
  

• 
  On most days, the automated call-back system appeared to be fairly reliable. The automated system appeared to be successful most of the time. Some of the contractors then attempted to make manual call-backs when the automated system failed. A small number of calls ere unsuccessful after both methods were used and these riders called-in to get their scheduled times.
  
• 
  The automated call-back system appeared to sometimes malfunction. This was reported to be rare, though, and past failures appeared to have been related to implementation or phone switchover issues. Kiessling and JV, which have a lower number of trips provided per day, reported that they are prepared to make manual call-backs when this happens. VTS and GLSS, the larger providers, indicated that they would rely on riders to call-in and request times when the automated system fails.
  
• 
  It was noted that in a system that requires call-backs, there will be issues reaching all riders regardless of the method used (automated or manual). In general, contractors felt the current automated system is an improvement over the prior manual call-back process.
  
• 
  It was recommended that all contractors be required to have a contingency plan to make call-backs manually should the automated system go down, or at least to adjust staffing plans to handle an increase in calls from riders. It also was noted that if manual call-backs are not made when the system occasionally is down, and additional staff is not added, relying on all riders to call-in to get their pick-up times would likely have a negative impact on evening and early morning dispatch operations and would likely result in long telephone hold times for all riders that evening and the following morning.
  
• 
  It was recommended that contractors should also be asked to keep a record of manual call-backs outcomes when manual call-backs are made for the small number of automated calls that are not successful, for TTY users, or for riders who have made special requests to have manual call-backs.
  

(Editor’s note: Data in the table shows the number of automatic call backs attempted, the number and percentage successfully completed, the number manually completed, and the number failed by service provider for December 2006 through February 2007. For each service provider, data is presented in the following order: Auto Call-Backs Attempted; Successful Automated Calls; Successful Manual Calls; Failed

December; 48,521; 43,211 (89.1%); 0; 1,368

January; 48,904; 44,528 (91.1%); 0; 1,028

February; 43,887; 39,928 (90.1%); 0; 1,045

GLSSGLSSGLSS;;;;

December; 35,225; 33,993 (96.5%); 139; 1,232

January; 35,598; 34,565 (97.1%); 172; 1,033

February; 32,991; 31,759 (96.3%); 154; 1,232

KiesslingKiessling;;;;

December; 25,861; 23,451 (90.1%); 82 ; 978

January; 27,236; 24,433 (89.7%); 62; 2,741

February; 24,541; 22,420 (91.4%); 30; 98

JVJV;;;;

December; 20,686; 18,061 (87.3%); 10; 309

January; 21,384; 18,041 (84.4%); 811 (3.8%); 360

February; 21,383; 17,570 (82.2%); 100 (.05%); 197

• 
  It was noted that a lot of pressure was placed on dispatch. As noted in the Scheduling section, afternoon capacity was an issue for some service providers. As a result, runs were tightly scheduled. A number of trips also were kept on the Wait List to be handled by dispatchers. In addition, same-day trip requests were encouraged in the rider material and these calls had to be handled by dispatchers. System-wide, about 10,877 same-day trip requests were received each month. About half of these (5,336 per month) were accommodated. Kiessling addressed this problem somewhat by ensuring that all trips were scheduled to runs and not left on the Wait List.
  
• 
  As noted later in the “Run Coverage” section of this report, dispatchers also were relied upon at some locations to handle trips on runs that were not covered when drivers are unscheduled “call-outs.” And, all of the regular duties of a paratransit dispatcher (managing scheduled runs and handling same day issues and incidents) also were the dispatchers’ responsibilities. It was observed that dispatch capacity was not adequate at some provider locations to meet all of these work demands.
  
• 
  It was noted that, even with additional dispatch capability, having so many things change on the day of service was a problem when the system was dealing with a set capacity. If there were flexible capacity that could be called on to help out during certain times (like the afternoon peak), these multiple same-day schedule changes might be able to be handled. Without flexible capacity, the set number of runs at times can be expected to be overloaded with changes and add-ons.
  
• 
  It was recommend that the MBTA work with each service provider to ensure adequate dispatch capacity. It also was recommended that the MBTA either: (1) work with the service providers to develop flexible capacity (extraboard drivers and vehicles, additional Road Supervisors, subcontracted taxi or van services), or: (2) work to limit and manage the number of same-day changes and add-ons.
  
• 
  In general, good no-show procedures appeared to be in-place and in use. Drivers were observed to consistently contact dispatchers when no-shows occurred. Dispatchers then consistently made attempts to locate riders and to ensure that drivers were at the correct location and had waited at least five minutes after the scheduled pickup time. Significant monitoring capability also existed with AVL and MDC technologies fully-implemented. Dispatchers appeared to be diligent in preparing time-stamped screen prints showing vehicle locations when no-shows occurred. The MBTA also appeared to be closely monitoring no-shows.
  
• 
  There was a potential issue with trips being cancelled or no-showed and rebooked in the dispatching process. This was done frequently when riders no-show and call later to reschedule their ride. It also was done when riders ask for same-day changes to their original schedules. It also could be done, though, if trips are running late and dispatchers transfer the trips to another vehicle or reschedule the trip. If this is done, initial trip information , such as the initial promised time is deleted (at least from standard reports). It then can be very difficult to determine if the trip was on-time. Careful monitoring in no-shows helped with this issue. It was recommended that a procedure for monitoring and tracking cancellations made in the dispatch process also be developed.
  
• 
  A potential accuracy issue when trips are rebooked or rescheduled in the dispatch process also was noted. As described in the “Scheduling” section above, the ADEPT system did not allow promised/scheduled times to be changed (which is a good protection). However, this therefore meant that to rebook a trip with a new scheduled time (such as after a no-show), the trip had to be rebooked and the trip information entered again. It was noted that this can cause inaccuracies and even missed trips if dispatchers become distracted during this rebooking process. It was recommended that the MBTA and service providers work with StrataGen to develop ways to more easily reschedule and rebook trips in the dispatch process while still protecting and preserving information about the original promised times.
  
• 
  It also was recommended that the MBTA encourage greater use of “Dispatch Notes” to carefully document same day service issues other than no-shows. This might include rider requests for earlier pickups, delays in rider boardings, and issues that contributed to trips being late or missed.
  
• 
  It was recommended that the MBTA and the service providers should resolve the interface problem between ADEPT and Mentor that is causing some trips to be incorrectly “performed” when no-shows are recorded.
  
• 
  It was recommended that the service providers explore creating “dispatch assistant” positions. These individuals would work in dispatch and would assist with handling calls from riders. Creating these positions would allow riders to be directed immediately to dispatch for same day service issues (eliminating a transfer from reservations and second hold time), and would allow full dispatchers to concentrate fully on managing runs.