Mobility-Driven and Energy-Efficient Deployment of Edge Data Centers in Urban Environments

in IEEE Transactions on Sustainable Computing (2021)

Multi-access Edge Computing (MEC) brings storage and computational capabilities at the edge of the network into so-called Edge Data Centers (EDCs) to better support low-latency applications. In this paper ... [more ▼]

Multi-access Edge Computing (MEC) brings storage and computational capabilities at the edge of the network into so-called Edge Data Centers (EDCs) to better support low-latency applications. In this paper, we tackle the problem of EDC deployment in urban environments. Previous research on mobile phone data has exposed a strong correlation between the demand for mobile communications and the urban tissue. For example, joint analysis of mobile data and vehicle traffic can be extrapolated to estimate demand for transportation and human activities, thereby inferring the land use of the area where such activities take place. Our work takes into account the mobility of citizens and their spatial patterns to estimate the optimal placement of MEC EDCs in urban environments, in order to minimize outages while guaranteeing energy-efficiency. This is achieved by modeling both the energy consumption attributed to network components (e.g., base stations) and computing components (e.g., servers). We propose and compare three heuristics and show that mobility-aware deployments achieve superior performance. The results are obtained with a custom-designed simulator able to operate over large-scale realistic urban environments. [less ▲]

Crowdsensed Data Learning-Driven Prediction of Local Businesses Attractiveness in Smart Cities

in IEEE Symposium on Computers and Communications (ISCC), Barcelona, Spain, 2019 (2019, July)

Urban planning typically relies on experience-based solutions and traditional methodologies to face urbanization issues and investigate the complex dynamics of cities. Recently, novel data-driven ... [more ▼]

Urban planning typically relies on experience-based solutions and traditional methodologies to face urbanization issues and investigate the complex dynamics of cities. Recently, novel data-driven approaches in urban computing have emerged for researchers and companies. They aim to address historical urbanization issues by exploiting sensing data gathered by mobile devices under the so-called mobile crowdsensing (MCS) paradigm. This work shows how to exploit sensing data to improve traditionally experience-based approaches for urban decisions. In particular, we apply widely known Machine Learning (ML) techniques to achieve highly accurate results in predicting categories of local businesses (LBs) (e.g., bars, restaurants), and their attractiveness in terms of classes of temporal demands (e.g., nightlife, business hours). The performance evaluation is conducted in Luxembourg city and the city of Munich with publicly available crowdsensed datasets. The results highlight that our approach does not only achieve high accuracy, but it also unveils important hidden features of the interaction of citizens and LBs. [less ▲]

Incorporating activity duration and scheduling utility into equilibrium-based Dynamic Traffic Assignment

in Transportation Research. Part B, Methodological (2018)

This paper deals with the problem of jointly modelling activity scheduling and duration within a Dynamic Traffic Assignment (DTA) problem framework. Although the last decades witnessed an intense effort ... [more ▼]

This paper deals with the problem of jointly modelling activity scheduling and duration within a Dynamic Traffic Assignment (DTA) problem framework. Although the last decades witnessed an intense effort in developing utility-based departure time choice models, relatively little has been done for understanding how the different assumptions on the utility model affect the model outputs. This problem is the main focus of this paper, which evaluates the effect of explicitly incorporating activity scheduling and duration within a generic user equilibrium DTA formulation. While using utility functions to model the positive component of the utility is a quite common procedure, the object of this paper is to show that a generic utility-based framework behaves as trip-based, activity-based, tour-based, or schedule-based if specific assumptions are specified. By establishing a set of properties, we quantify the amount of utility lost due to traffic congestion and how this affects activity (re-)scheduling and duration decisions. This allows predicting the effect of using a different assumption on the evolution of the transport system – and more specifically the departure time choice model. Conclusions support the idea that, under specific conditions, complex user behaviour can be approximated through a simplified model, and that the ratio between utility at origin and destination can be used to identify systematic biases within an existing DTA model – such as anticipating the rush hour. We also propose a novel utility function suited for modelling different activities, which can be used for modelling activities with a different duration. The mathematical model used to evaluate the effect of scheduling and duration into the equilibrium-based Dynamic Traffic Assignment is a simple bottleneck model. While this model has been recently re-formulated in order to capture the interaction between morning/evening commute, this paper further generalizes it in order to account for all type of activities. [less ▲]

Dynamic Origin-Destination Matrix Estimation with Interacting Demand Patterns

Doctoral thesis (2018)

It has become very fashionable to talk about Mobility as a Service, multimodal transport networks, electrified and green vehicles, and sustainable transportation in general. Nowadays, the transportation ... [more ▼]

It has become very fashionable to talk about Mobility as a Service, multimodal transport networks, electrified and green vehicles, and sustainable transportation in general. Nowadays, the transportation field is exploring new angles to solve mobility issues, applying concepts such as using machine learning techniques to profile user behaviour. While for many years “traffic pressure” and “congestion phenomena” were the most established keywords, there is now a widespread body of research pointing out how new technologies alone will solve most of these issues. One of the main reasons for this change of direction is that earlier approaches have been proven to be more “fair” than “effective” in tackling mobility issues. The main limitation was probably to rely on simple assumptions, such as in-elastic mobility travel demand (car users will stick to their choice), when modelling travel behaviour. However, while these assumptions were questionable twenty years ago, they simply do not hold in today's society. While it is still true that high-income people usually own a car, the concept of urban mobility evolved. First, new generations are likely to buy a car ten-twenty years later than their parents. Second, in many cases, users can choose options that are more effective by combining different transport modes. Wealthy people might decide to live next to their working place or to the city centre, rather than to buy a car. Thus, it becomes clear that to understand the evolution of the mobility demand we need to question some of these assumptions. While data can help in understanding this societal transformation, we argue in this dissertation that they cannot be considered as the sole source of information for the decision maker. Although data have been there for many years, congestion levels are increasing, meaning that data alone cannot solve the problem. Although successful in many case studies, data-driven approaches have the limitation of being capable of modelling only what they observed in the past. If there is no record of a specific event, then the model will simply provide a biased information. In this manuscript we point out that both elements – data and model – are equally relevant to represent the evolution of a transport system, and specifically how important is to consider the heterogeneity of the mobility demand within the modelling framework in order to fully exploit the available data. In this manuscript, we focus on the so-called Dynamic Demand Estimation Problem (DODE), which is the problem of estimating the mobility demand patterns that are more likely to best fit all the available traffic data. While this dissertation still focuses on car-users, we stress that the activity based structure of the demand needs to be explicitly represented in order to capture the evolution of a transport system. While data show a picture of the reality, such as how many people are travelling on a certain road segment or even along a certain path, this information represents a coarse aggregation of different individuals sharing a common resource (i.e. the infrastructure). However, the traffic flow is composed of different users with different trip purposes, meaning they react differently to a certain event. If we shut down a road from one day to another, commuting and not commuting demand will react in a different way. The same concept holds when dealing with different weather conditions, which also lead to a different demand pattern with respect to the typical one. This dissertation presents different frameworks to solve the DODE, which explicitly focus on the estimation of the mobility demand when dealing with typical and atypical user behaviour. Although the approach still focuses on a single mode of transport (car-users), the proposed formulation includes the generalized travel cost within the optimization framework. This key element allows accounting for the departure time choice and, in principle, it can be extended to the mode choice in future work. The methodologies presented in this thesis have been tested with a “state of the practice” dynamic traffic assignment model. Results suggest that the models can be used for real-life networks, but also that more efficient algorithm should be considered for practical implementations in order to unleash the full potential of this new approach. [less ▲]

Evaluating the reliability of the Utility-Based Dynamic OD Estimation on Large Networks

Scientific Conference (2018, January)

Using Passive Data Collection Methods to Learn Complex Mobility Patterns: An Exploratory Analysis

Scientific Conference (2018)

Effectiveness of the Two-Step Dynamic Demand Estimation model on large networks

in Proceedings of 2017 5th IEEE International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS) (2017, June 28)

In this paper, the authors present a Two-Step approach that sequentially adjusts generation and distribution values of the (dynamic) OD matrix. While the proposed methodology already provided excellent ... [more ▼]

In this paper, the authors present a Two-Step approach that sequentially adjusts generation and distribution values of the (dynamic) OD matrix. While the proposed methodology already provided excellent results for updating demand flows on a motorway, the aim of this paper is to validate this conclusion on a real network: Luxembourg City. This network represents the typical middle-sized European city in terms of network dimension. Moreover, Luxembourg City has the typical structure of a metropolitan area, composed of a city centre, ring, and suburb areas. An innovative element of this paper is to use mobile network data to create a time-dependent profile of the generated demand inside and outside the ring. To support the claim that the model is ready for practical implementation, it is interfaced with PTV Visum, one of the most widely adopted software tools for traffic analysis. Results of these experiments provide a solid empirical ground in order to further develop this model and to understand if its assumptions hold for urban scenarios. [less ▲]

Generating Macroscopic, Purpose-Dependent Production Factors Through Monte Carlo Sampling Techniques

in Transportation Research Procedia (2017)

Effects of incorporating activity duration and scheduling utility on the equilibrium-based dynamic traffic assignment

Scientific Conference (2016, July)

Systematic analysis of global and local control policies

in Periodica Polytechnica Transportation Engineering (2015)