For Whom The Empty Bus Runs : How can DRT reduce the cost inefficiency of fixed-route bus?

It is said that if DRT (Demand Responsive Transit) replaces the existing fixed-route bus, users' waiting time is reduced and the budget of public transportation agency is saved.

This argument is not always correct. If most of the bus routes overlap with the passengers' travel corridors, running a bus rapid transit (BRT) with a short headway is much more efficient than running a DRT.

However, if the population density is low or passengers' travel origins and destinations are spatially distributed, DRT is a cost-effective way compared to a fixed route bus.

But to what extent is DRT more cost-effective than fixed route bus? Replacing a fixed route bus with DRT incurs additional costs, such as operating a call center for fleet hailing. We also have to pay extra to the DRT operation company (i.e. Studio Galilei ). Can DRT save budget compared to a fixed route bus when all these costs are included?

This article presents an example of the Republic of Korea. The 11 bus lines in the analysis were converted to DRT in 2023. Of course, these 11 lines had a very small number of users at the time of operation as fixed routes, so they had a very large proportion of subsidy among the total operating costs. Therefore, these bus lines are expected to have high efficiency improvements after switching to DRT. Were the results as expected?

First of all, a decision of the measurement index is required for comparative analysis. We used the transport cost per person (cost/person (USD)) and the cost (cost/(person*km) (USD)) paid to transport 1 km of one passenger as comparative measures. Each bus route was not named specifically. Data collected for 9 days were used for each bus and DRT.

First, let's look at the comparison based on (cost/person). We first analyzed the operating cost of the fixed route bus. Out of a total of 11 lines, the most expensive bus line (line 3) spent 119.1 USD transporting one passenger. On the other hand, the bus route, which was the lowest cost, spent 10.6 USD of transportation cost. Since the budget required to operate one bus per day does not differ significantly from region to region, this cost is almost determined by the number of passengers using the bus line. In the case of the route that had the most expensive transportation cost, 3 buses were operated, and the average number of passengers per day was only 13 pax.

Since the transition from bus to DRT, transportation costs have decreased significantly. Route 3, which required 119 USD for single passenger delivery, only needs to 22 USD, a 74% reduction in transportation costs. The operating cost of DRT also includes both the cost paid to the platform company (Studio Galilei ) for service operation and the cost of operating the call center additionally. The average cost of single-person transportation on 11 bus lines was 45.8 USD, but the average cost of single-person delivery by DRT decreased to 12.4 USD.

Of course, the biggest reason for the decrease in (cost/person) is the increase in passengers. The number of daily passengers on all 11 lines was 927 when buses were operated, but increased to 1,635 after switching to DRT. However, the researchers at Studio Galilei discovered another important fact while comparing the cost of buses and DRTs. Haven't you already found an important difference through above picture?

Another big change after the bus was converted to DRT is that the (cost/person) difference between 11 bus lines was greatly reduced. At the time of operation as a fixed bus route, the standard deviation of the (cost/person) of 11 lines was 31.3, but the standard deviation decreased to 4.5 after DRT runs. In other words, when it was operated as a fixed route bus, the difference in transportation cost among 11 bus lines was large, but after changing to DRT, this difference decreased significantly.

It can be seen that the fixed route buses before DRT conversion were incurring large cost losses due to inefficient detour routes and operation schedules. However, it can be seen that the inefficiency caused by incorrect route and fixed operation schedules has significantly decreased as it was converted to dynamic DRT. In other words, it was confirmed that the cost of unoptimized route and operation schedules can be decreased through DRT operation.

If so, are the results of the comparative analysis of (cost/(person*km)) the same to the previous one? The figure below is a spider web graph using (person*km/cost) of 11 bus lines. From the results, it can be seen that the standard deviation between lines has decreased significantly after DRT conversion like (cost/person).

However, in the (cost/(person*km)) analysis, it is also found that it is very different from the previous analysis. In the figure above, in the case of bus line 7, after DRT conversion, (cost/(person*km)) increased from 2.1 USD to 2.7 USD.

Bus line 7 has two peculiarities. Once converted from the bus to DRT, the number of passengers did not increase significantly. During bus operation, the average number of daily passengers on bus line 7 was 81, and after conversion to DRT, it was still 87. And when operating as a fixed route bus, the detour of the route was considerably long. Therefore the passenger's travel distance was also long. But after DRT conversion, most passengers are delivered on the shortest path. In result, the passenger's travel distance decreased. In other words, the total amount of (person*km) decreased as the unnecessary travel distance was shortened after DRT conversion.

Considering that unnecessary detours of the fixed route bus are not desirable, it can be seen that there may be such distortion when using (cost/(person*km)) as an evaluation index when analyzing the business effect after the DRT conversion of the bus.

There is one more important fact besides that. In the case of Bus line 7, when it was operated as a fixed route bus, long-distance travel services were provided to nearby communities. Although this bus line took a very long time due to many detours, not a small number of passengers took this route to a distant nearby community. However, after the DRT conversion, the average travel distance was also reduced because only internal community services were provided. The decrease in (cost/(person*km)) shows that the number of residents who needed such long-distance travel was not small. Therefore, if a decrease in (cost/(person*km)) is found after switching from the bus to DRT, it will be necessary to review whether there is a change in life satisfaction due to the shrink of socio-economic activity area of the residents.

In this article, we looked at what changes the overall budget is when the inefficient fixed route bus is converted to the real-time operating DRT. And we also looked at what measurement index should be used for the comparison.

Most of the results were as expected, but some were unexpected. The most important finding among them is to prove that real-time dynamic DRT eliminates the inefficiency of fixed-line buses. Mathematically, this result is valid. Fixed route and fixed operation schedules play the role of constraint in vehicle operation, and in mathematics, the solution to the optimization problem under the constraint equations derives an inefficient solution compared to the unconstrained optimization problem. It is mathematically natural that DRT, which can serve vehicles to passengers without being constrained by fixed route and fixed operation schedules, is more efficient than fixed route buses.

For this reason, it is correct to convert an inefficient fixed route bus to DRT. As a public transportation expert, I would like to ask the question 'For whom the empty bus runs' quoting the classic 'For whom the bell rolls'. For whom is the bus in the rural area running, which costs 80 USD to transport a passenger?

All the inefficiencies maintained through government assistance that can be deleted through technology must now disappear. This is because the budget is all spent on citizens' wallets. Leaving service that can be improved without reducing the budget that can be reduced should no longer be allowed in the realm of welfare. It is to take away the rights of other citizens who should receive welfare benefits.

In the next article, I would like to introduce another example of budget reduction that occurred in the DRT business in Korea. I hope interested readers will read this as well. And researchers from all over the world who want to work on this together are welcome at any time. Studio Galilei or contact Hyunmyung Kim .