Home

The FITS project aims to develop a prototype platform to explore the feasibility of a virtual transport market place to more efficiently match existing demand and supply for transport services in rural areas.

Background

One of the principal causes of failure in existing vehicle brokerage solutions is that they require the introduction of a single trusted third party to whom transport offers and requirements are sent, and which solves the scheduling problem. Advances in AI planning and scheduling could be utilised to address the scalability issues inherent here, but such refinements do not address the key need to decentralise decision-making. This is not to say that matchmaking of potential transport suppliers to consumers is not essential, but information from such a service should inform rather than determine the transport options for customers. The decentralisation of decision-making enables transport users and providers to retain their autonomy while collaborating with others when beneficial. Intelligent agent technology offers a means to facilitate this collaborative decision-making. Taking, for example, the transport user in a rural area, the inputs to their decision-making will include transport/mobility goals, individual constraints that capture the (normative) restrictions on what transport options they would consider acceptable, and (potentially) transport resources at their disposal (e.g. a private car). The ability to query potential transport options (from suppliers such as taxi firms, bus operators, a neighbour with a private car, etc.) is important, but will not be the sole source of information. Other sources include reports of strikes, weather warnings, accident reports, etc. Such information may quite readily be made available in machine-useable form that can be utilised by an agent supporting a user. Transport information may also be available from non-traditional sources such as fellow transport users.

Given this context, how can intelligent agents support the decision-making and information seeking activities of individuals and groups to best meet their transport needs? There are many research challenges here. Information from various sources may be more or less trusted by the user; how can evidence from such varied sources be combined so that an agent can best advise a user of the transport options that meet their needs with justifications for these options. There may be no option that meets the user’s goals given their normative constraints; how can an agent best advise the user on the implications of the constraints they impose? Transport providers, or individuals offering shared travel may wish to push information to potential customers/collaborators, or established social networks of individuals in a specific area may wish to collaborate in deciding whether they can share, for example, a taxi to and from a main bus/train stop to catch specific services; how can the dialogue among these stakeholders be effectively managed, and, possibly automated, by the agents representing them? And, related to this, how can we ensure that agents are trusted to operate according to the information privacy requirements of these stakeholders?

The technological approach that is proposed here is the use of intelligent agents that act on behalf of the stakeholders within the system, and the use of argumentation mechanisms to identify options open to users, weighing the evidence for desirability of each option given a model of the user’s priorities , and to drive dialogue among agents in aiding users to solve their individual (or collective) transport goals. Existing research in agent support for transport resource management has typically been focussed on the provider. Our vision is to explore both the efficient use of limited transport resources, but also to support the passenger in the decision-making process.

Objectives

• Rapid review of recent international experience of flexible transport services, real-time scheduling and intelligent autonomous agents
• Establish, in the context of a suitable case study, user requirements (including issues of “trust”) of key stakeholders (e.g. government and local authority decision makers, commissioning bodies for transport services, transport operators (including taxis), subsidised transport client groups and general public)
• Generate data to estimate demand and supply schedules for taxi services and examine the role of incentives to determine how drivers respond to changes in earning opportunities
• Specification of a platform for supporting autonomous agents acting on behalf of stakeholders to solve their individual (or collective) transport goals
• Development of prototype platform
• Testing of the prototype platform in selected case study area(s) with co-operation of local authority partners to establish validity of approach, perhaps assisted by use of the dot.rural Digibus
• Assessment of future uptake of integrated flexible transport services in rural areas with an extended stakeholder base

---