Customized Mass Transit: Vision and Research Agenda (2008)
Unedited archival documents establishing a system architecture for multi-modal transit consolidation.
Author’s Historical Note
The texts below are unedited copies of documents that were published on my university webpage in 2008. Originally developed in 2006, the architecture was initially slated for commercial deployment in collaboration with a veteran high-tech executive. After those commercial plans stalled, I published the model as an open R&D framework to ensure the concepts entered the public domain.
Overview of Customized Mass Transit (CMT)
CMT is an integrated software platform that synthesizes public transit, safe scheduling, and door-to-transit shuttles into a single, seamless travel solution. By combining mobile GPS, total cost optimization, and dynamic dispatch, CMT established the multi-modal consolidation framework now operating commercially as the combination of public transit, Uber’s Pool, and Last Mile logistics.
Structure of This Document
- Part 1 (De Facto Executive Summary): A concise, two-paragraph overview of the CMT system architecture, user convenience factors, and macro-economic benefits.
- Part 2 (Full R&D Proposal): A more detailed 5-page proposal outlining the multi-carrier coordination system, its technological enablers (GPS, web, mobile), interface requirements, and the underlying stochastic economic models needed for trip scheduling and dynamic pricing.
Professional inquiries can be directed to my LinkedIn profile.
© 2008 Dan Trietsch. All rights reserved.
_____________________________________________
Customized Mass Transit:
To encourage commuters and other day-travelers to utilize privately- or publicly-owned mass transit modes (such as vans, buses, light-rail, trains and ferries), we must make it possible to obtain door-to-door service and make it easier to plan and pay for trips. Customized Mass Transit (CMT) is a web-based solution that coordinates trips and utilizes multiple modes. The system relies on the existing and currently-developing electronic infrastructure; e.g., it would make extensive use of cell phones and their built-in GPS capabilities. At its core, it requires advanced scheduling algorithms of the type we have been developing in the last few years.
Get Dan Trietsch’s stories in your inbox
Join Medium for free to get updates from this writer.
A typical trip might start by a dial-a-ride shuttle taking a commuter to from home to the train station (alongside others). After a train ride, the commuter might ride a bus or another shuttle to the office. Payment might be coordinated by the CMT service provider (which is just one way to make the service highly profitable). Advantages of the service to the commuter include convenience (no stop-and-go driving, looking for parking, etc), cost savings (no need to purchase that third car, less expenses on gas, etc) and, depending on the modes involved, even time savings (by using fast car-pool and bus lanes, fast trains, etc). Advantages to service providers include higher utilization of equipment. Advantages to society include energy savings, reduced congestion and the ability to use electricity and modern energy-saving solutions (such as hybrids) more extensively and earlier than would be the case if we had to wait for the current fleet of conventional vehicles to be replaced.
_____________________________________________
Customized Mass Transit: Vision and R&D Agenda
Overview: I propose a web-based platform to provide customized transit trip planning and booking for commuters and other short trips. Using the product will allow the commuter to leave his/her car at home and improve capacity utilization and service frequency offered by transit service providers. Users will have the option to receive door-to-door service and yet be able to utilize fast and efficient mass-transit modes. One way to deploy such a platform is already in existence and involves basic scheduling for commuters who use public transit services in certain markets (e.g., the German train system, public transit in Paris). However, the fully developed system should be capable of coordinating services from multiple public and private carriers. The service provider (i.e., the user of the proposed software product) can also collect the fare on behalf of the carriers. Furthermore, the service provider can also charter transit capacity to provide part of the service; e.g., cabs and shuttles may be hired to provide door-to-door service and the software will coordinate their trips. Service providers may be private or public, but the main feature is coordinating the services of multiple independent carriers. Benefits to passengers include avoiding the need to drive, cost savings, potential time savings by using public transit fast lanes and avoiding the need to park. Benefits to society include reducing congestion, increasing utilization of public transit vehicles, energy savings, facilitating the introduction of new public transit services and the ability to shift passengers to environmentally cleaner modes faster than would be the case if private vehicles continue to be the main transit mode.
General Discussion: Essentially I am proposing a platform that will support a self-developing flexible multi-modal transit system. Parts of the physical transit system can be owned by either public or private entities, and none of them is required to provide a complete travel solution. Instead, complete trip solutions are to be created by a consolidator using an advanced software application that combines services from the various carriers. Similar services have existed in the long distance passenger transportation and cargo shipping environments for a long time. That is, travel agents and cargo consolidators provide a similar service for more expensive movements. There are also systems that combine various modes for commuters (such as rail, light rail and bus) but typically these are simply parts of a single public transit system. In contrast, the purpose of this proposal is to facilitate coordinating services provided by multiple carriers in a seamless manner as far as the passenger is concerned. To this end, when comparing transit to long distance passenger transportation or cargo transportation, transporting commuters and local passengers involves much more detail and yet the value of a typical trip is much lower. Therefore, until now, it would not be economical to create such a system. As costs of optimization and computing drop, however, and computing capacities increase, it is now feasible to design such a system. Furthermore, there are also pressures coming to a boil currently that make such a development especially important. The proposed system can be started immediately and yet it can use new services in future that do not exist today. For example, if an entrepreneur decides to provide ferry service between two points, A and B, with a given frequency, a given price and a given reliability, the system can utilize that new service for all trips for which it makes economic sense to do so. From the point of view of the entrepreneur, this facility increases the total demand for the new service and thus makes it more economically viable than otherwise. Since the new service is not likely to provide a complete solution for many passengers, this is a very important catalyst. The existence of such a system removes a significant barrier to entry that would exist if the same entrepreneur would also be required to provide service to and from points A and B for all potential passengers. Most importantly, the system should be attractive enough to encourage present and past commuters to leave their car at home, or perhaps not even own a car in the first place. Short term rental vehicles — rented by the hour mostly in central city districts (CBD) — can also be incorporated in such a system. Conceptually, and especially if the service is public, even car pools can be coordinated on an ad hoc basis. Ideally, with such a system in place, even well-to-do commuters should prefer public transit over driving. Rail, sea and air passenger transportation always included First Class service to cater for the rich. Under the vision promoted here, the market is likely to continue offering basic services as well as options that are more attractive to the affluent. Indeed, the system should offer a menu of alternatives for each trip varying in terms of fare, time consumption and the service level (defined as the probability of completing the trip by an agreed deadline). The same customer may select a cheaper and slower service on one day, and an expensive fast service on another. The system will schedule the service and — in case a shuttle is involved — instruct the customer when to be ready in a way that balances the cost of waiting of the customer with the cost of waiting of the shuttle and other commuters already in it (each of whom may have a different time value). For this purpose, stochastic economic balance models are required.
Background: Since the second half of the 20th Century, the dominant passenger transportation mode in America and much of the rest of the industrialized world has been the private automobile. Due to various influences, including illegal activities by major automobile, oil and tire giants, the role of public transit has declined tremendously during that period, so much so that large part of the public transit system are now very unattractive to most passengers: they often serve mainly the disenfranchised. Emerging economies, including China and India, show signs of following the same blueprint: private car ownership in these countries is increasing exponentially. But this system is now under three major sources of pressure to change: (i) congestion; (ii) the energy crunch; and (iii) a growing consensus that climate change is associated with fossil-fuel based energy consumption of the type that current automobiles rely on (and will continue to rely on for the next couple of decades at least). Solutions involving new automotive technologies and new clean energy sources are being developed and they will certainly help alleviate the energy crunch and possibly the climate change aspects in future (the question whether this will be too late for climate change or not is outside the scope of this proposal). There is little doubt, however, that such technological solutions will not have a major effect in the next decade at the very least: at best we can turn the tide in that period. Furthermore, such solutions will do nothing to alleviate road congestion. In fact, the climate change issue may make it progressively difficult politically to develop the road infrastructure to the extent necessary to alleviate congestion.
On the one hand, there is a strong argument that the solution must be based on public transit modes and not on the private car. On the other hand, it might be impossible to turn the clock back and ask commuters to give up the door-to-door convenience associated with the current preferred mode. The question is whether we can design a transportation system that will allow passengers to maintain their desired convenience level and yet not rely on their own car. Better still, can we design a system that can achieve that and yet be composed of both privately and publicly owned elements? If so, then market forces could be relied on to provide the necessary solutions. But such a benign state can only emerge if commuters will be able to plan their own trip with negligible effort. To that end, we need a coordination system that can do such planning for commuters in a manner that is transparent to them. The purpose of this R&D proposal is to design a coordination system and suggest what infrastructure might be needed to achieve the desired result. The system only involves proven technologies and relies on four enablers that information technology has already provided: (i) powerful computers; (ii) the world wide web; (iii) the GPS system; and (iv) mobile phones (with built in GPS capabilities). But because it was not until very recently that all these enablers became available, there is room for further research and new synthesis.
The importance of the enablers is that they can provide the power to process millions of travel requirements and facilitate the communications between the passenger and the system. Indeed, if we consider the case of cargo transportation (excluding mail), it has a similar volume but by far less items to coordinate. As a result, it has been possible to coordinate the shipment of cargo across several modes (e.g., truck-rail-ship-rail-truck) with individual carriers coordinating their activities in a way that is transparent to the customer and virtually seamless. Travel agents provide similar services to long distance passengers but, again, that is economically feasible because there are fewer yet much more expensive trips (in comparison to transit). In principle, however, passenger transit requires a similar structure, with trips potentially utilizing several modes (e.g., bike-bus-train-bus), and yet commuters should be able to use the system without having to personally coordinate all the details and with simple combined billing for the trip. Otherwise, the complexity of planning is likely to push passengers to less efficient single-mode solutions.
The Role of the Free Market in the System Development: One of the fundamental premises is that the system must not require enforcing central coordination on individual service providers. Instead, it should be designed in such a way that market forces will be allowed to play their beneficial role. Furthermore, it should allow for any conceivable portfolio of available services, as well as services that may become available in future. And yet, from the point of view of a typical commuter, the system must provide seamless connections and look like a single virtual system. Thus, the system’s main role is to coordinate travel over multiple modes and multiple carriers, charge passengers for whole trips and reimburse carriers for their share in these trips. Another premise is that the system must provide tangible benefits to the public as a whole in the form of savings of energy (especially if it involves greenhouse gases and pollutants) and a reduction or even elimination of the need to increase the capacity of the highway infrastructure. By reducing congestion and using car pool lanes, the system should also be effective in terms of providing timely service. It is assumed that enough of these savings will be transferred to the customers and thus provide an economical incentive to use the system. Yet another premise, however, is that commuters should not be forced to walk or bike if they prefer
motorized transportation from door-to-door. Any attempt to do that will greatly reduce the willingness and in some cases the ability of some commuters to patronize the system.
Dial-a-Ride: The scheduling of individual trips through the existing transportation network is central to the system. It may be called the brain of the system. But such a system also requires the development of dial-a-ride options. Dial-a-ride is a way to provide efficient door-to-door system by shuttle, where several passengers share a minibus. Passengers order the service and the service provider combines and coordinates their trips. It is also desirable to introduce frequent connections on main arteries both leading to- and from the CBD and within the CBD. The dial-a-ride service would actually pick passengers from home and deliver them either to their final destinations or to an intermediary service (such as a train or a frequently-served artery leading to the CBD). If a train or a ferry is utilized, another shuttle may take the passenger from the other train or ferry station to the destination itself.
Research and Development: Developing such a system involves two main parts: interface and algorithms. The interface should coordinate all the necessary details and communicate with the passengers and service providers. The algorithmic part involves designing (and improving) the algorithms used for scheduling decisions within the system. Research questions that this system requires answers to are numerous. Focusing on operations research issues, a prime example is solving the dial-a-ride stochastic scheduling problem to serve heterogeneous passengers with different time-values and service level requirements. Another example is optimizing the capacities of various parts of the system; on the one hand, it is quite likely that market forces will achieve much of this optimization by themselves, but on the other hand, it is still a government responsibility in most countries to plan the infrastructure, and such a system influences the correct design of that infrastructure and could benefit from good models. All these examples involve stochastic balance; i.e., no component in the system should be allowed to limit it consistently but rather a resource should be the system bottleneck with a frequency proportional to the cost of increasing its availability. Stochastic balance also helps plan individual trips, per the general principles of safe scheduling (i.e., devising schedules with optimal safety time).
Other research issues have to do with estimating the costs and benefits that such a system involves. Such costs have several facets: energy, pollution, congestion, passenger time and outright monetary expenditures are involved. Using such costs as input, research is also required for price-setting in such a manner that passengers will be rewarded for making desirable choices and taxed for overtaxing the environment and the infrastructure. It is also important to research the effect of such a system: How much fuel can be saved? How helpful might that be? How much time can be saved to passengers, if any? How to allocate expensive technologies to vehicles (e.g., vehicles that are only used during rush hour might justify less investment)?
