Periodic Reporting for period 1 - NETWORKS (NETWORKS Doctoral Program)
Reporting period: 2020-09-01 to 2022-08-31
Networks for communication, transportation, finance and energy form the backbone of modern society. Reliable and efficient network infrastructures are therefore of enormous economic and social value, and their importance will only increase in the coming years. The complexity of networks and the processes running on them is increasing rapidly as well, which regularly leads to unpredictable and uncontrollable behaviour, and poses severe threats. Networks not only play a vital role for society, but they have also evolved into rich sources of data. These data are crucial for understanding structural properties and optimizing functional performance of large-scale networks. To deal effectively with the uncertainty, variation, unpredictability, size and complexity inherent in complex networks, we must develop radically new ways of thinking. The NETWORKS program uniquely combines stochastics (to model and understand large-scale networks and to predict network growth and network processes) with algorithmics (to control and optimize networks and network processes in the best possible way). The symbiosis of these two areas, one rooted in mathematics and the other in computer science, is what allows NETWORKS to make a decisive contribution to the advancement of network science. This symbiosis is underpinned by two paradigm shifts that have taken place in modern science: “randomness is everywhere” and “the need for fast algorithms”.
There is an urgent need for a generation of researchers who know how to deal with contemporary and emerging networks that are inherently stochastic in nature, and at the same time know how to design effective and well understood decision and optimization algorithms. The NETWORKS doctoral program will train 14 such researchers, who will gain the necessary scientific knowledge and professional skills for a successful career inside and outside academia, and who will contribute to the challenges posed by complex networks, by performing top-level research in an international environment. The main objectives of the NETWORKS doctoral program are:
- To bring together expertise from two different disciplines, namely stochastics (a subarea of mathematics) and algorithmics (a subarea of computer science), and train our ESRs by performing world-class research in, and at the interface between, these disciplines. Our multi-disciplinary approach is essential for complex network problems, since the performance of network algorithms is influenced by the random nature of network processes.
- To attract talented young researchers and offer them the best possible training and career perspectives. The extensive training program developed by NETWORKS includes scientific training in stochastics and algorithmics, and a broad range of key professional skills and personal development.
- To stimulate transnational and intersectoral mobility and knowledge transfer. NETWORKS will attract excellent young researchers from outside the Netherlands. Our large international network will provide them with excellent additional opportunities for transnational mobility. Intersectoral mobility and knowledge transfer are ensured through secondments at industrial and other non-academic partners of NETWORKS.
There is an urgent need for a generation of researchers who know how to deal with contemporary and emerging networks that are inherently stochastic in nature, and at the same time know how to design effective and well understood decision and optimization algorithms. The NETWORKS doctoral program will train 14 such researchers, who will gain the necessary scientific knowledge and professional skills for a successful career inside and outside academia, and who will contribute to the challenges posed by complex networks, by performing top-level research in an international environment. The main objectives of the NETWORKS doctoral program are:
- To bring together expertise from two different disciplines, namely stochastics (a subarea of mathematics) and algorithmics (a subarea of computer science), and train our ESRs by performing world-class research in, and at the interface between, these disciplines. Our multi-disciplinary approach is essential for complex network problems, since the performance of network algorithms is influenced by the random nature of network processes.
- To attract talented young researchers and offer them the best possible training and career perspectives. The extensive training program developed by NETWORKS includes scientific training in stochastics and algorithmics, and a broad range of key professional skills and personal development.
- To stimulate transnational and intersectoral mobility and knowledge transfer. NETWORKS will attract excellent young researchers from outside the Netherlands. Our large international network will provide them with excellent additional opportunities for transnational mobility. Intersectoral mobility and knowledge transfer are ensured through secondments at industrial and other non-academic partners of NETWORKS.
NETWORKS has successfully attracted 14 talented young researchers, from a variety of countries inside and outside Europe. They started working on challenging research projects, supervised and coached by senior researchers in the NETWORKS team, who are all renowned experts in their field. The supervisors come from mathematics (with expertise in queuing theory, applied probability, random graphs, large deviation theory, combinatorial optimization, and more) and computer science (with expertise in algorithms for spatial data, graph algorithms, quantum algorithms, and more). The many training activities organized in NETWORKS further provide the young researchers with an excellent and multi-disciplinary training. This is enhanced by the professional-skills training that they receive. Thus, at the end of their PhD studies, the young researchers are well prepared for a leading role in Europe's knowledge economy.
NETWORKS performs ground-breaking research on stochastic and algorithmic network problems. The research is organised around eight themes.
1. Approximate and exact network methods
Algorithmic network problems networks are often NP-hard, meaning that no fast algorithms exist that solve these problems optimally on all possible instances. We study two approaches to deal with this. One is to develop fast algorithms that compute almost optimal solutions. Another is to exploit that real-life instances often have structural properties that can be leveraged to obtain fast algorithms.
2. Spatial networks
Many real-world networks are spatial: nodes have a location in space and edges are defined by physical connections or geographic proximity between the nodes. We study how and in which situations it is possible to exploit the geometry of the network, to obtain better solutions to network problems.
3. Quantum network algorithms
Quantum computers are based on the laws of quantum mechanics. NETWORKS focuses on quantum software for networks problems. A key question is which computational problems can be solved significantly faster on quantum computers, and which are still very hard.
4. Dynamics of networks
Networks evolve over time, in a way that is typically closely related to their functionality. Within NETWORKS we develop and analyse random-graph models and we investigate which models are best for which applications.
5. Dynamics on networks
Network functionality can often be described in terms of stochastic processes taking place on networks. Within NETWORKS we investigate how the behaviour of stochastic network processes is affected by the irregular structure of the network, in particular, the presence of “hubs”.
6. Transportation networks
The efficient usage of road, railway and other transportation networks poses many mathematical challenges. Research within NETWORKS deals both with structure-related issues (planning and dimensioning of transportation and traffic networks) and with the operations on existing networks (routing and other traffic management mechanisms).
7. Communication networks
Communication networks need to be designed to consistently achieve high levels of performance and reliability, and yet be cost-effective to operate. We study how processes on networks (such as the spreading of viruses or fake news) evolve and can be controlled, and how to construct and control communication networks to maximize efficiency.
8. Energy networks
The shift towards renewable energy sources such as wind and solar energy is causing a significant variability in supply to electricity networks. As a result, supply and demand may no longer match at any given time. Our research aims at getter a better grip on this by developing and analysing novel mathematical models for energy networks.
In addition, the NETWORKS project will also strengthen the European human capital base through delivering fourteen experienced researchers in the vital field of network science.
1. Approximate and exact network methods
Algorithmic network problems networks are often NP-hard, meaning that no fast algorithms exist that solve these problems optimally on all possible instances. We study two approaches to deal with this. One is to develop fast algorithms that compute almost optimal solutions. Another is to exploit that real-life instances often have structural properties that can be leveraged to obtain fast algorithms.
2. Spatial networks
Many real-world networks are spatial: nodes have a location in space and edges are defined by physical connections or geographic proximity between the nodes. We study how and in which situations it is possible to exploit the geometry of the network, to obtain better solutions to network problems.
3. Quantum network algorithms
Quantum computers are based on the laws of quantum mechanics. NETWORKS focuses on quantum software for networks problems. A key question is which computational problems can be solved significantly faster on quantum computers, and which are still very hard.
4. Dynamics of networks
Networks evolve over time, in a way that is typically closely related to their functionality. Within NETWORKS we develop and analyse random-graph models and we investigate which models are best for which applications.
5. Dynamics on networks
Network functionality can often be described in terms of stochastic processes taking place on networks. Within NETWORKS we investigate how the behaviour of stochastic network processes is affected by the irregular structure of the network, in particular, the presence of “hubs”.
6. Transportation networks
The efficient usage of road, railway and other transportation networks poses many mathematical challenges. Research within NETWORKS deals both with structure-related issues (planning and dimensioning of transportation and traffic networks) and with the operations on existing networks (routing and other traffic management mechanisms).
7. Communication networks
Communication networks need to be designed to consistently achieve high levels of performance and reliability, and yet be cost-effective to operate. We study how processes on networks (such as the spreading of viruses or fake news) evolve and can be controlled, and how to construct and control communication networks to maximize efficiency.
8. Energy networks
The shift towards renewable energy sources such as wind and solar energy is causing a significant variability in supply to electricity networks. As a result, supply and demand may no longer match at any given time. Our research aims at getter a better grip on this by developing and analysing novel mathematical models for energy networks.
In addition, the NETWORKS project will also strengthen the European human capital base through delivering fourteen experienced researchers in the vital field of network science.