One of the first changes brought on by COVID-19 was a huge drop in public transportation ridership and regulatory demands that service be reduced. Agencies scrambled to make these service changes, rushing in days through a planning and scheduling process that typically takes weeks and quite a lot of scheduling know-how. According to our estimates, the lion’s share of U.S. agencies have already finished the creation of these service reduction contingency plans.
At the same time, agencies began grappling with questions of bus operator safety, questions that we believe will persist when things begin to return to normal, hopefully soon.
At Optibus, we’re offering transit agencies a free license of our platform so they can create contingency plans. In the process, we learned some lessons, and we’d like to share them with the industry.
Three Ways to Reduce Service
Agencies took three approaches to making the service changes. The first approach was to reduce routes or service according to ridership decreases. A typical example can be commuter services where the passengers opt to take their private car, for reasons of social distancing and lower levels of congestion. This involves making ad-hoc changes to routes or eliminating some altogether. The second approach was taking the weekend schedule (a Saturday or Sunday schedule) and applying it to all days of the week. This approach preserved the existing runcuts of the weekend service and required creating a new roster. The third approach was to make changes across a broad range of service metrics, affecting coverage, frequency, span or the length of routes. This requires making changes in routes or timetables, creating new vehicle and crew schedules to support them and then creating a roster. In ordinary times and on legacy systems, this type of service change typically takes weeks - in this case it had to happen in a day or two.
From what we’ve heard from agencies, the easiest option was the weekend service option, since it only required rostering changes and no new runcuts or blocks. It was also easier to communicate to the public, because it involved referring to a pre-existing weekend schedule. However, the resulting service isn’t perfect, since it doesn’t take into consideration passengers who may be still commuting to work, and who possibly are emergency workers, nurses, etc. In this second option, rostering, if manual (we estimate that 70 percent of U.S. agencies don’t perform roster optimization but rather work manually) can take a day or two, while with optimization, it can take an hour or two and result in less roster lines.
Bus Operator Safety and Absenteeism
Initially, when COVID-19 requests began coming in, the regulatory focus was dealing with contingencies of a reduction of 10 or 25 percent of bus operators, due to issues ranging from operator fear of working with the public to inability to work because of the virus or quarantines. While we’ve looked at creating runcuts where the main consideration is driver shortages, reducing the maximum number of drivers by such high percentages won’t necessarily deliver a schedule that makes sense. That’s why most agencies focused on using the weekend service.
In any case, driver protection measures require that optimization be used to ensure a minimal amount of rosters, to preserve the health of bus operators by exposing as few as possible to the public, even with measures such as no fare or backdoor boarding. This also helps deal with potential absenteeism caused by the pandemic.
Other safety measures include creating crew and vehicle schedules that involve frequent cleaning and disinfection of buses and the driver area. An example can be added breaks for disinfection or allocating more time for changeovers, so that the driver area can be disinfected as it passes from one driver to the next. Other options, which we haven’t seen, but may be relevant in case of a gradual return to ordinary service, are disallowing changeovers in order to ensure that each driver is present only in one bus. Eliminating changeovers is bound to hurt efficiency and cost, but can work in the short term, where service is reduced anyway.
Post COVID-19 - An Opportunity to Reshape Rosters
The need to quickly generate rosters made agencies use roster optimization for the first time, instead of the cafeteria-style rostering they had in the past. There are several downsides to cafeteria-style rostering even during non-COVID-19 service operation. The obvious ones are lack of control of overtime and rosters that are less efficient cost-wise. Yet the most compelling reason is that optimized rosters create better rosters for everyone, since more quality rosters are created and fairness is ensured. This eliminates have and have-nots in rosters and creates an opportunity to reshape roster practices post COVID-19.
What Next?
Looking at data from public transit in China, which is slowly returning to normal, ridership decreases persist even when the pandemic has been beat out of cities. If a similar behaviour persists in the U.S., we expect to see the same scramble to plan and reschedule many service changes since it will be a gradual return to normal. Hopefully, this downtime will also be used to improve underlying scheduling practices such as rostering, changeovers and more.
----------