Expert in Intelligent (Transport) Systems


My research interests are motivated by the understanding that our world is increasingly automated and interconnected. The number of devices that we rely on in conducting our daily lives is growing, as is their level of intelligence, autonomy and collaboration. Ensuring that such devices and their networks operate reliably, efficiently and according to our expectations has become a major challenge.

A rigorous framework through which this challenge can be tackled in a principled way is provided by multi-agent systems, a scientific discipline at the intersection of artificial intelligence, distributed computing, economics and social sciences. Multi-agent systems provide models, theories and algorithms for describing, analysing, designing and managing distributed systems composed of large numbers of interacting autonomous entities, termed agents.

Out of many topics which nowadays comprise the field of multi-agent systems, I'm mostly interested in agent-based simulation modelling and multi-agent planning and resource allocation, and, importantly, in their application to intelligent transport systems. 

Next Generation Multi-modal Journey Planning

The growing number of transport options available raises the importance of tools that empower travellers and allow them to tailor journey itineraries to their specific needs. Existing journey planners support such requirements only to a limited degree and further research on data models and algorithms is needed to support rich multi-modal journey planning.

I'm interested in extending routing and journey planning algorithms to support the full spectrum of mobility services available in modern cities, combining individual and collective, fixed-schedule as well on-demand means of transport while taking into account individual user preferences and availability of transport services and resources.

Many of these goals I've already achieved, primarily as the result of large collaborative R&D projects SUPERHUB and MyWay. Among the more interesting ones is the innovative meta-planning approach to journey planner integration which allows rapidly building and deploying a fully multi-modal journey planner on top of existing mode-specific journey planners. Another key result are novel algorithms for multicriteria bicycle routing in complex urban environments, which has been partly integrated in our bicycle router prototype.

Algorithmic Transport System Accessibility Analysis

Understanding how different areas of a city or a region are accessible by different means of transport is essential for evidence-based mobility policy management and planning. I'm exploring how advanced graph search algorithms, combined with detailed data about the structure of a transport network, can be used to calculate various transport accessibility metrics including travel times, number of interchanges and service frequencies. One result of this research is the on-line Transport Analyser that allows the user to interactively explore urban accessibility by public transport, car and bike in an easy-to-use web application.

Intelligent Marketplaces for On-Demand Transport Services

Much of the inefficiency in existing transport systems stems from the lack of coordination between the actors in the system. Due to the multi-party, partially competitive nature of the problem, multi-agent models and techniques are particularly well positioned to deliver solutions for improving such coordination and thus the efficiency of transport systems.

I'm therefore exploring how market-based mechanisms, in particular auctions, can be used to better plan and schedule the use of limited transport services and resources. I'm particularly interested in developing the methodology for optimally designing market-based mechanisms for real-time on-demand mobility services exemplified by companies such as Uber or Liftago.

Fully Agent-based Transport Simulation

Simulation modelling has long been an essential tool for transport planning and management. Enabled by the ever increasing computational capabilities, traffic and transport simulation approaches have been becoming increasingly agent-oriented.

My personal interests in this area revolves around the following topics:

  • Simulation testbeds for on-demand transport systems – I explore how agent-based simulation models can be used as testbeds for analysing on-demand mobility and logistics services, including real-time ride sharing, transportation network services or next-generation taxi services. One result of this research is the open-source Flexible Mobility Services Testbed.
  • Fully agent-based transport modelling – I explore how transport systems can be modelled as genuinely multi-agent systems, composed of autonomous agents with continuous, asynchronous control modules and the ability to interact freely with the environment and other agents. Such a fully agent-based approach – which is not supported by existing platforms – improves the modularity and extensibility of transport simulations and reduces constraints on the type of decision models that can be simulated (e.g. within-the-day replanning). See also the RODOS project.

Past Research Interests

In the past, I conducted research on computational game theory, adversarial reasoning techniques and contract-based systems. Although these areas remain of interest of me, I'm no longer actively pursuing  research on those topics.