Taxibus Case Study

Let us examine how a taxibus fleet compares to existing forms of transport. We shall take Greater London as our case study, but our analysis can be applied to most major cities. During each 24 hour period in central and outer London, around 25 million passenger journeys take place. These divide modally as follows.

9 million   Private Car (as Drivers)

6 million   Private Car (as Passengers in above)

4 million   Bus

2 million   London Underground (Metro)

2 million   Walking

1 million   Train (Surface Rail)

Source: Department for Transport. Figures have been rounded.

Other modes of travel in Greater London (motorcycle, taxi cab, mini cab, bicycle) are fairly negligible in comparison. The relative importance of these transport modes is somewhat reversed as far as peak-hour travel to and from central London is concerned: morning peak hour (7 am to 10 am) passenger journeys into central London divide modally as follows.

0.5 million   Train (Surface Rail)

0.4 million   London Underground (Metro)

0.1 million   Private Car (Either as Driver or Passenger)

0.05 million  Bus

Source: Department for Transport. Figures have been rounded.

How would a London-based taxibus fleet compare? Consider a fleet of just 10 thousand taxibus vehicles. Assume an average vehicle occupancy of 6 passengers, and a typical passenger journey time of about 20 minutes (since the average vehicle speed in London is 20 mph and the average journey length is 7 miles). A quick calculation then shows that such a taxibus fleet would complete 180 thousand passenger journeys every hour. In a 24 hour period, the fleet could in principle handle 4.3 million door-to-door passenger journeys, which is an absolutely formidable number.

We can compare this taxibus fleet to London's 6 thousand buses, which carry 4 million passengers each day, mainly in double-decker vehicles having a capacity of around 80 passengers. Another point of reference is London's 20 thousand licensed taxi cabs which, throughout every 24 hour period, manage to transport 0.5 million passengers.

How will the fleet of 10 thousand taxibuses cope, theoretically at least, with the 7 am to 10 am rush hour commute from the suburbs to central London? Assuming, during these rush times, an average taxibus occupancy of 8 passengers and an average commuter journey length of say 40 minutes, then a simple calculation shows that the taxibus fleet will manage to transport over 0.36 million commuters during these three hours. This is a very impressive figure, almost equalling the performance of the entire London Underground system during the same time period.

What about the operating profit and the running costs of the taxibus fleet? The major running costs will arise from driver wages: this fleet of 10 thousand taxibuses will require around 30 thousand drivers working in 8 hour shifts to keep these vehicles on the road 24 hours a day, 7 days a week (based on only a reduced taxibus service running at night). Assume each driver costs a total of £25,000 (US $44,000) per year for salary and other costs, then the wage bill will amount to £750 million (US $1.3 billion) annually. Fuel costs for this fleet can be calculated at £300 million (US $520 million) a year, assuming the fleet uses diesel fuel charged at UK prices, with each taxibus giving around 20 miles per gallon. Note that fuel costs will be much lower in the US, where fuel is less expensive.

Further costs will be incurred from vehicle servicing and maintenance, and the expenditure of running the IGT computer system. Telecommunications costs should be minimal, since only low volumes of data are transmitted. Driver training costs will also be low: a standard driver's licence will suffice since most of the taxibus vehicles will be not much larger than a regular minibus or 'people carrier' type of motorcar. Taxibus drivers will not need to have knowledge of the street layouts, since electronic street navigation information is provided by the IGT system at all times.

What about the sales revenue generated? Setting the taxibus passenger fare at a modest 15 pence (US 26 cents) per mile, and assuming an average vehicle occupancy of 6 passengers, with an average vehicle speed of 20 miles per hour, then the daily (24 hour) revenue collected by the fleet will be £4.3 million (US $7.5 million). This adds up to nearly £1.6 billion (US $2.8 billion) a year, which should easily cover the running costs of the taxibus fleet.

Regarding the capital cost of introducing such a taxibus fleet: most of the investment will arise from buying taxibus vehicles; the computer technology that runs IGT is relatively cheap. Assume each taxibus vehicle costs £30,000 (US $52,000), which is the typical current price of a multi-purpose vehicle or minibus: it will then require a capital investment of £300 million (US $520 million) to purchase a fleet of 10 thousand taxibus vehicles. Not only is this a modest figure by transport budget standards, but this amount might even be accommodated within the annual profits generated by such a fleet.

It is interesting to examine the figures for a larger taxibus fleet. Suppose the fleet is increased to 30 thousand taxibuses. This requires a capital investment of £900 million (US $1.6 billion) to buy the vehicles, a team of 90 thousand drivers to keep them moving, £2.25 billion (US $3.9 billion) a year in driver salaries, and £900 million (US $1.6 billion) a year on fuel; such fleet will generate an annual revenue of £4.7 billion (US $8.2 billion), and provide almost 13 million door-to-door passenger journeys each day - a figure that is close to the daily total of passenger journeys made by private car in London. Thus a fleet of 30 thousand taxibuses can do more or less the same job as London's 2.3 million privately-owned cars. We could therefore dispense with many of these cars. This possibility is not just a utopian vision: it is a glimpse at what cities of the future will be like. The taxibus has the potential to transform urban life and the city landscape.

Does it make economic sense for a Londoner to sell his car and travel primarily by taxibus? Let us consider this question. Ownership of the average car can be calculated at around £10 (US $17) a day inclusive of fuel, road tax, insurance, servicing, maintenance, and vehicle depreciation, but exclusive of parking costs. The average vehicle covers a distance of less than 30 miles each day with an average occupancy of 1.4 travellers: this equates to a cost of 24 pence (US 42 cents) per person per mile, excluding any parking fees. Comparing this to the 15 pence (US 26 cents) per mile taxibus cost, this analysis suggests that it does make good economic sense to sell the car and adopt the taxibus, although the precise economics will greatly depend on individual circumstances. Families with children may not be so keen to sell the car, since the car becomes more economically viable when there are more travellers. A special discount on taxibus fares might be offered to family groups in order to make the taxibus more financially appealing to families.

The ultimate objective of the taxibus is not to eliminate the private car, but to seduce people away from the car by providing a taxibus transportation network that the public will come to view as a much more convenient and convivial way to travel. The taxibus aims to compete with - and beat - the private car in terms of transport excellence; winning people over just by the superb service IGT offers.

Note: the above transportation and financial performance analysis of the taxibus is based on an assumed average vehicle occupancy of just 6 passengers; if higher average vehicle occupancies are achieved, say 10 passengers per vehicle for example, then all the above performance figures will be proportionally improved.

 

Three Minute Response Time

The success of the taxibus as an everyday means of transport will crucially depend on the time a prospective passenger has to wait for a taxibus to arrive to pick him up. To compete with the 'jump in and go' immediacy of the private car, it is considered that a taxibus must appear within three minutes of a passenger making a journey request. Is such a rapid response feasible? Some simple analysis can answer this question. Again we shall take Greater London as our case study (but the conclusions of our analysis are applicable to other towns and cities).

The total area of Greater London (the city centre plus the suburbs) is around 610 square miles. Assuming we have a fleet of 10 thousand taxibuses more or less evenly spread across this area, this gives an average vehicle density of 16 taxibuses per square mile. The UK's Department for Transport statistics indicate that the average road speed in London is around 20 mph. In order for a taxibus to arrive within three minutes of the passenger submitting his journey request, the responding taxibus must be situated no further than one mile away (since at the average road speed of 20 mph, it takes exactly three minutes to travel one mile). How many taxibuses are there within a one mile radius of the typical waiting passenger? The answer is simply the area of a one mile radius circle (which is 3.142 square miles) multiplied by the taxibus vehicle density per square mile. This gives 16 x 3.142 = 50 taxibuses. This is a fair number of vehicles, and since these 50 taxibuses will have a wide variety of itineraries, there is a high chance that one or more will be travelling in a direction compatible to the passenger's desired destination. (This probability also depends on the closeness of the passenger's destination: the closer the destination, the higher the chance of finding a taxibus with a compatible itinerary. Note that the average journey length in London is just 7 miles, so most destinations are in fact quite nearby.)

This simplified analysis is no substitute for a more in-depth mathematical modelling of taxibus response times, but it does strongly suggest that a three-minute response is eminently feasible.

The other important consideration in studying the feasibility of a three-minute passenger pick-up response relates to the computer processing power necessary to perform optimal routing within this timescale. The mathematics of optimal routing often involves calculations that consume a lot of computer power (the computer processing time can be huge, since it is typically proportional to the exponential of the fleet size). For this reason, most Demand Responsive Transport (DRT) systems find it hard to calculate optimal routes even given 24 hours notice, and the very the fastest-responding DRT systems require half an hour's notice, in order that the computer can have time to determine the optimal routes for passenger pick-up and delivery. So how can IGT perform the almost instantaneous route optimisation that is necessary for on-the-fly routing and a passenger pick-up that is within three minutes? On first analysis, this seems like a mission impossible.

Paradoxically, however, the tight constraint of a three-minute response actually makes taxibus route optimisation much easier. This is simply because this rapid response greatly reduces the number of routing possibilities in the computation. When a prospective passenger submits his journey request to the IGT system, only a relatively small number of taxibuses out of the whole fleet will be able to reach him within three minutes, so obviously the system just needs to examine the itineraries of these close-proximity taxibuses in order to determine the optimum vehicle to carry the passenger. Mathematicians term this process of cutting out irrelevant possibilities from the calculation as 'pruning'. There are many further pruning possibilities even within this subset of close-proximity taxibuses.

So finding an optimal route for an IGT taxibus can be done pretty rapidly: not much computer power is really required. It is very important to appreciate this double advantage of the three-minute response. Not only does this attractive feature put IGT way ahead of all rival DRT systems, but in addition, IGT's three minute response greatly reduces the computer time required to perform the optimal routing calculation. In this sense, the three-minute response is self-enabling.

Finally, note that without the cellular telephone and its ability to send, receive and display text data, the three minute response - and IGT in general - would be less feasible. Sure, you can use other communications means for ordering a taxibus, but the cellular phone is obviously extremely convenient, ubiquitous, fast and simple to use. The advent of cellular telephony is a vital enabling technology for IGT.