Networking Transportation Networkers


Bioevolution of transport networks with slime mould

Andrew Adamatzky
University of the West of England, Bristol UK

Slime mould Physarum polycephalum is a monstrous single cell well known for its task-solving abilities — solves computational geometry and logical problems, navigates robots and generates music. The slime mould could also build motorways, highways and expressways. It is used to analyse transport networks of Africa, Australia, Belgium, Brazil, Canada, China, Germany, Iberia, Italy, Malaysia, Mexico, the Netherlands, UK and USA. the largest cities are represented by oat flakes and the slime mould is inoculated in a capital. When all oat flakes are covered by the slime mould, the structure of the protoplasmic networks formed are analyzed. In the laboratory experiments and theoretical analyses, intriguing country-specific properties of the motorway networks are uncovered and compared with the man-made and slime mould networks. They are studied as proximity graphs, leading to hierarchies of complexity and bio-rationality of the motorways.


The formation, evolution and computational behaviour of Physarum-inspired multi-agent transport networks

Jeff Jones
University of the West of England, Bristol UK

True slime mould Physarum polycephalum is a giant single-celled amoeboid organism that grows, moves, forages and feeds via morphological adaptation of its protoplasmic nutrient transport network. Despite having no nervous system, special senses, skeletal structure, or organised musculature Physarum exhibits complex biological behaviour, can perform difficult computing and robotics tasks and has been the subject of extensive research in the past decade. Much of this research is to try to discover how the complex behaviour in Physarum can emerge from such simple component parts. In this talk we describe a multi-agent model of Physarum, adopting a virtual material approach to the formation, growth and adaptation of emergent transport networks.

The model, comprising a population of mobile chemotactic particles, represents both the structure and flux within the Physarum plasmodium, using auto-catalytic coupling of particles based on reaction-diffusion dynamics. The particle collective undergoes spontaneous network self-assembly and exhibits second-order minimisation properties approximating surface tension dynamics including the observation of Plateau angles, and von Neumann's law. The biological patterns of the Physarum plasmodium are reproduced, including the initial formation and subsequent morphological adaptation of the protoplasmic transport network, and the growth and foraging behaviour towards attractants, and away from hazards. Nutrient dependent growth patterns of the Physarum plasmodium are demonstrated and a parsimonious mechanism explaining the complex biological behaviour is suggested, whereby the plasmodium exploits pre-existing chemoattractant gradients, and also modifies these gradients in a two-way interaction with the environment. 

The dynamical evolution of the transport network structure as it undergoes adaptation is investigated and the computational properties of the emergent transport networks are described, including results on the construction and minimisation of proximity graphs, approximation of Voronoi diagrams (and hybrid variants), and the formation of convex hulls and concave hulls. Can we control the evolution of these transport networks for computational purposes? A method of dynamically controlling the evolution of the networks for combinatorial optimisation is described, along with a simpler approach based on material shrinkage.

To conclude, we find that Physarum networks, and their computational models, emphasise the role of dynamical adaptation in transport networks, with network structure adapting to a changing environment. This may provide opportunities – and challenges – to the study of transport networks in general.


Self-Control of Traffic Lights

Stefan Lämmer
Technische Universität Dresden, Dresden Germany 

Based on fluid-dynamic and car-following simulations of traffic flows in urban networks, we propose a self-organization approach to traffic light control. It assumes a priority-based control of traffic lights by the vehicle flows themselves, taking into account short-sighted anticipation of vehicle flows and platoons. The considered local interactions lead to emergent coordination patterns such as 'green waves' and achieve an efficient, decentralized traffic light control. We show that critical queue spillbacks can be avoided by an inflow-regulating traffic light control that does not serve more vehicles than subsequent roads can accommodate. In this way, vehicular queues only build up within designated areas of the roads segments, whereas upstream intersections remain fully accessible for non-affected flow directions. In consequence, congested parts of the network are relieved from some traffic, whereas remaining capacities on surrounding roads are utilized. The result is a considerable reduction not only in the average travel times, but also of their variation.


Elementary processes governing the evolution of road networks

Vincenzo Nicosia
School of Mathematical Sciences, Queen Mary University of London, London UK

The evolution of transportation networks in general and of street networks in particular is often considered a good proxy to characterise the urbanisation of an area. In this talk we show the results of the empirical analysis of a unique data set regarding almost 200 years of evolution of the road network in a large area located north of Milan (Italy). We find that the urbanisation of the area under study has dramatically changed the structure and organisation of the road network, e.g. through the progressive homogenisation of cell shapes, while, at the same time, high–centrality roads which constitute the backbone of the urban network have remained pretty stable over time, confirming the importance of historical paths. By using some simple metrics introduced in complex networks analysis, we show quantitatively that the growth of the network is governed by two elementary processes: (i) ‘densification’, corresponding to an increase in the local density of roads around existing urban centres and (ii) ‘exploration’, whereby new roads trigger the spatial evolution of the urbanisation front. The empirical identification of such simple elementary mechanisms suggests the existence of general, simple properties of urbanisation and opens new directions for its modelling and quantitative description.

Ref: E. Strano, V. Nicosia, V. Latora, S. Porta, M. Barthelemy, Scientific Reports, vol 2, 296, doi:10.1038/srep00296 (2012). 


Foreland analysis using complex networks: last results for general cargo and containerized transport

Carlos Pais Montes
University of A Coruña, A Coruña Spain

Last decade has been of a great development in containerized transport patterns, and high value added goods travel across world maritime routes responding quickly to the currently high volatile geographical demand conditions.

So big hubs have experienced a logical shift in their loading/unloading/trading dynamics, but also medium and small size ports, like the ones placed in our Galicia-North Portugal region (mainly A Coruña, Ferrol, Marin, Vigo and Leixões), which are faced to a entire change in world transportation network structures.

This new global reconfiguration of transportation paths might have positive effects for territorial growth if ports are able to react properly not only improving their infrastructures (building of new deep draught terminals, automation of calling processes, new regulatory and law environments) but also performing innovative new markets-seeking operations (targeting which geographical areas are playing crucial roles at a regional scale).

Within this conceptual framework, the researcher’s team of the Institute of Maritime Studies (IMS) in University of A Coruña, have been working in the use of Graph Theory to describe 2007-2011 worldwide maritime transport activity, and also in the definition of new methodologies (complex networks based) aiming to extract solid and accurate measures of containerized and general cargo trade at a regional level (foreland descriptions).

The results of the IMS complex networks researching program show several worldwide patterns which will require further investigation: a) the process of containerized concentration in few hubs with a strong network of subsidiary ports (port regionalization); b) the validity and good health of general cargo (non containerized) transport modes; c) the demand attraction of the new emergent economies; d) the new reconfiguration of US East and West coasts trade connections due to the next opening of the new Panama Canal lanes; e) the force of the Mediterranean containerized terminals (despite deep crisis affecting most of their countries); f) the paper of Atlantic Arc ports (United Kingdom, Ireland, France, Spain and Portugal Atlantic shore) in the new emergent trans-Atlantic activity expected.


Networks, morphology and land use location: a case study of the London region

Peter Ferguson
Centre for Advanced Spatial Analysis, University College London, London UK 

In this presentation I will review the use of street network accessibility measures in a planning and urban design context, providing an overview of their practical application, strengths and limitations in professional practice. I then present a set of alternative network potential measures of the London region that while primarily suitable for the masterplanning of local street layouts, land use distributions and building massing, begin to incorporate the impact of the wider transport network on the potential of local neighbourhoods to support commercial activity.


What metros have to teach us, from complexity to centrality

Sybil Derrible
University of Illinois at Chicago, Chicago USA

Beyond their function to mass transport people, metros have a myriad of insights to teach us. In particular, metros constantly experience self-organization processes. Indeed, despite being built in places varying in demography and culture, despite being built over long periods of time, and despite being built to satisfy the specific needs of a city, metros all around the world possess common features. The main objective of this talk is to: (1) introduce a methodology to analyze metros as networks, and (2) present an analysis of the topology of 33 metros in the world. In particular for (1), carefully selecting and defining the sets of nodes and links is desirable to output meaningful results (only the termini and transfer stations are taken into account in this research). To tackle (2), a variety of metrics are used to analyze how metros evolve with network size. In particular, metros can be characterized according to their state, form and structure. Moreover, metros also tend to have complex network properties. Overall, the presence of these topological properties is relevant and they can markedly impact ridership. Finally, in more recent work, the network centrality of metros (and specifically betweenness centrality) is studied yet again showing clear patterns with network size.


Modeling the Dynamics of Public Transport Networks – Theory and Applications

Oded Cats
KTH Royal Institute of Technology, Stockholm Sweden

Public transport systems exercise complex dynamics and evolve through the interactions among numerous agents. Moreover, the importance of transfers, multi-modality and limited spatial and temporal availability has implications on public transport networks connectivity and vulnerability. The dynamics of public transport evolve through the complex interactions between traffic conditions, system operations and travelers’ demand. An agent-based public transport simulation model was developed in order to mimic the emergence of global spontaneous order from numerous inter-dependent local decisions. In this seminar, the framework of such a public transport operations and assignment model will be presented. The model involves the integration of several modules including traffic simulation, public transport operations and control, dynamic path choice model and real-time information generator. Applications in various domains will be presented. In particular, applications in the domain of network vulnerability, links centrality and capacity and the impacts of mitigation measures will be discussed.