Network slicing is one the key features of software-defined networks (SDNs) and can be used in next-generation communication networks. Admission control of network slices is the basis of providing the heterogeneous quality-of-service performance guarantee and maximizing the optimization objectives of the network operator. Various admission control mechanisms have been proposed in the literature, including those based on traffic forecasting. However, recurrent neural network-based probabilistic forecasting models have not been given thorough consideration for slice admission control. In this study, the network slicing scheme design problem is formulated mathematically, with an equivalent formulation of the constrained bandwidth sharing scheme. Then, a DeepAR-based slice admission control mechanism is proposed for sequential decision making for network slice requests in SDN, with the support of the SDN controller. An improved variant is further proposed with a closed-loop parameter update mechanism. The experiments based on real-world historical traffic data validate the effectiveness of the proposed mechanisms, with metrics including revenue, resource reservation and utilization ratios, and service admission ratio.

As an alternative to traditional meteorological methods, rain attenuation in satellite-to-Earth microwave communication signals has been used for rainfall reconstruction in recent years. In this article, the existing 2-D rain field reconstruction problem is extended to a 3-D scenario by leveraging the low Earth orbit satellite system. A compressive sensing approach is further proposed to solve the 3-D rain field reconstruction problem. The Starlink system is used as a reference, and two synthetic rain events near the Great Barrier Reef in Australia, which are generated from the weather research and forecasting model, are used to evaluate the reconstruction performance. Simulation results show that the compressive sensing approach performs better than both the traditional least squares and the least absolute shrinkage and selection operator approaches.

Since great redundancy of telemetry data of spacecraft, telemetry data compression is a good solution for the limited bandwidth and contact wireless links. It is important to obtain accurate data characteristic firstly. State-of-the-art machine learning methods work well on data mining and pattern recognition under conditions of the given test data set, which could be used as the available tools for post-event data processing and analysis, such as trend forecasting and outlier detection, but they have not provided the proper solution from the source on-board. In this paper, four base classes of the telemetry data are suggested and studied through the time series feature and information entropy analysis, then a new on-board lightweight self-learning algorithm named Classification Probability calculation - Window Step optimization (CP-WS) is proposed to obtain the class features and make the decision of each single parameter from the continuous discrete telemetry time series. Simulation results show that, our algorithm correctly classifies the simulation and real mission data into the appropriate base class with advantages of high classification accuracy as 100% and adaptive computational complexity from $O(L^{2})$ to $O(L)$ , which could be used in satellite on-board data compression for space-to-ground transmission, especially for the deep space explorers to save important status with less on-board storage space.

Traffic forecasting has been an active research field in recent decades, and with the development of deep-learning technologies, researchers are trying to utilize deep learning to achieve tremendous improvements in traffic forecasting, as it has been seen in other research areas, such as speech recognition and image classification. In this study, we summarize recent works in which deep-learning methods were applied for geospatial data-based traffic forecasting problems. Based on the insights from previous works, we further propose a deep-learning framework, which transforms geospatial data to images, and then utilizes the state-of-the-art deep-learning methodologies such as Convolutional Neural Network (CNN) and residual networks. To demonstrate the simplicity and effectiveness of our framework, we present a formulation of the New York taxi pick-up/drop-off forecasting problem, and show that our framework significantly outperforms traditional methods, including Historical Average (HA) and AutoRegressive Integrated Moving Average (ARIMA).

In this paper, we introduce a bipartite matching model for matching markets with dynamic arrivals and departures. Different from classical models with a finite-time horizon, our model has a long-time horizon with infinite vertices. In our model, the matching goal is to maximize the ratio of matched vertices, i.e., matched ratio. We define two types of online algorithms, i.e., Greedy and Patient, analyze their performance with evaluation metrics of both upper bounds and competitive ratios, and conduct extensive simulations to validate our analysis. To further simulate the real situation, we extend our model with the user’s strategic behavior and prove the existence of a specific Nash equilibrium under a differentiated matching mechanism.

In this paper, we propose a large-scale nationwide ridesharing system named CountryRoads, which is designed to resolve the surge of homeward-bound persons during holiday seasons. CountryRoads matches drivers who are willing to share their spare seats with passengers who are unable to obtain public transportation for long-distance trips. The system utilizes a variety of user interfaces, including text messages, mobile websites, and smart-phone applications to enable the participation of a variety of people. Further, it implements an online greedy-matching algorithm to solve the ridesharing problem in order to propose matches between drivers and passengers.

We deploy and evaluate CountryRoads during the Chinese Spring Festival travel season (Chunyun period), when all forms of transportation are stressed or near saturation, and when many lower-income individuals fail to reserve seats for rail or flight trips. We present our design choices, analytics, and lessons learned from our multi-year deployment in 2012, 2014, and 2015. The improvements over the years resulted in up to 17,272 users, 4777 ridesharing tuples, and an eventual success rate of 23.2% in 2015. This is a significant result considering that most of our participants use CountryRoads as a last resort in response to public transportation shortages.

To gain insight into how transportation network companies, such as Uber and Didi, impact the taxi industry, we conduct a multi-period analysis of taxi drivers' behaviors, based on GPS trajectory data collected from three time periods in Beijing, i.e., November 2012, November 2014, and November 2015. We extract both passenger-delivery and passenger-searching trip information from GPS trajectories and compare the spatial, temporal, densification, and poolability properties of taxi trips in different time periods. Our results reveal that the taxi industry was adversely influenced by the competition between transportation network companies; as compared with that of 2012, the average passenger-delivery trip number per day per taxi dropped by 18.08% and the average daily profit per taxi dropped by 19.29% in the year 2015, respectively. We also compare passenger-searching strategies, passenger-delivery strategies, and service area preferences between taxi drivers with top and bottom efficiency in different time periods. We find that compared with drivers with lower efficiency, drivers with high efficiency tend to search locally, have a higher delivery speed, and serve more often within the inner part of Beijing.

Double-apping happens when passengers or drivers use multiple smartphone-based e-hailing applications. While it has been observed in practice, its effects on the e-hailing service have not yet been explored by any previous study. In this paper, we give the first systematic study of the phenomenon of double-apping and clarify its definition in a smartphone-based e-hailing market model with two ridesharing companies. We also define four cruising modes for drivers, i.e., stationary, random, gravity, and historical. To study the effects of the double-apping and compare different cruising modes, we conduct extensive simulations based on real taxi GPS trajectories. Our results demonstrate that: 1) the network effect exists in the e-hailing market, which makes the driver ratio of the market leader at equilibrium grows faster with higher market share; 2) the gravity and historical cruising modes outperform the other two modes, in terms of assigned ratio, dispatch distance, order number, and driver's profit; 3) passenger's and driver's double-apping have the opposite effects for a ridesharing company; and 4) the ridesharing companies are in a Prisoner's Dilemma, where two-sided double-apping cannot be achieved and a unilateral action of allowing driver's double-apping brings no benefit to the ridesharing company itself.

To gain insight into how transportation network companies like Uber and Didi impact the taxi industry, we conduct a multi-period analysis of taxi drivers' behaviors, based on GPS trajectories collected from three time periods in Beijing. We extract both passenger-delivery and passenger-searching trip information from GPS trajectories and evaluate taxi drivers' working conditions. Our results reveal that the taxi industry was adversely influenced by the competition between transportation network companies, as compared to that of 2012, the average passenger-delivery trip number per day per taxi dropped by 18.08% and the average daily profit per taxi dropped by 19.29% in year 2015, respectively. We also compare passenger-searching strategies, passenger-delivery strategies, and service area preferences between taxi drivers with top and bottom efficiencies in different time periods. We find that compared to drivers with lower efficiencies, drivers with high efficiencies tend to search locally, have a faster delivery speed, and serve within the central part of Beijing.

In this study, we propose a crowd sensing framework with the existence of execution uncertainty and a given budget. Our framework consists of three stages: Task Selection, Task Allocation, and Payment. Within each stage, we define the design problems and give a preliminary solution with desirable theoretic properties.

The Chinese Spring Festival travel season (Chunyun) has been called the largest annual human migration in the world with approximately 3.6 billion trips in 2014. Understandably, all forms of transportation are stressed at or near saturation during this period and many people, especially lower-income individuals fail to get to their destinations. We present CountryRoads, a ridesharing system to address the transportation shortage during Chunyun. The CountryRoads system collects users' route information, and matches drivers and passengers as an online bipartite matching problem based on the proximity of the passenger's origin and destination and the route of the driver. The system is evaluated during four Chunyun periods from 2012 to 2015 with up to 17272 users and forms 4777 ridesharing tuples in 2015.