-
Physics-Informed Graph Neural Network for Dynamic Reconfiguration of Power Systems
Authors:
Jules Authier,
Rabab Haider,
Anuradha Annaswamy,
Florian Dorfler
Abstract:
To maintain a reliable grid we need fast decision-making algorithms for complex problems like Dynamic Reconfiguration (DyR). DyR optimizes distribution grid switch settings in real-time to minimize grid losses and dispatches resources to supply loads with available generation. DyR is a mixed-integer problem and can be computationally intractable to solve for large grids and at fast timescales. We…
▽ More
To maintain a reliable grid we need fast decision-making algorithms for complex problems like Dynamic Reconfiguration (DyR). DyR optimizes distribution grid switch settings in real-time to minimize grid losses and dispatches resources to supply loads with available generation. DyR is a mixed-integer problem and can be computationally intractable to solve for large grids and at fast timescales. We propose GraPhyR, a Physics-Informed Graph Neural Network (GNNs) framework tailored for DyR. We incorporate essential operational and connectivity constraints directly within the GNN framework and train it end-to-end. Our results show that GraPhyR is able to learn to optimize the DyR task.
△ Less
Submitted 2 April, 2024; v1 submitted 1 October, 2023;
originally announced October 2023.
-
Capacitor Optimization in Power Distribution Networks Using Numerical Computation Techniques
Authors:
Jordan R. Keuseman,
Chad M. Smutzer,
Clifton R Haider,
Barry K. Gilbert
Abstract:
This paper presents a power distribution network (PDN) decoupling capacitor optimization application with three primary goals: reduction of solution times for large networks, development of flexible network scoring routines, and a concentration strictly on achieving the best network performance. Example optimizations are performed using broadband models of a printed circuit board (PCB), a chip-pac…
▽ More
This paper presents a power distribution network (PDN) decoupling capacitor optimization application with three primary goals: reduction of solution times for large networks, development of flexible network scoring routines, and a concentration strictly on achieving the best network performance. Example optimizations are performed using broadband models of a printed circuit board (PCB), a chip-package, on-die networks, and candidate capacitors. A novel worst-case time-domain optimization technique is presented as an alternative to the traditional frequency-domain approach. The trade-offs and criteria for scoring the computed network are presented. The output is a recommended set of capacitors which can then be applied to the product design.
△ Less
Submitted 17 January, 2023;
originally announced May 2023.
-
Plated-Through-Hole Via Design Specifications for 112G Serial Links
Authors:
Michael J. Degerstrom,
Chad M. Smutzer,
Richard B. Ericson,
Clifton R Haider,
Barry K. Gilbert
Abstract:
An earlier study of a high layer-count test board using plated-through-hole (PTH) vias and a limited quantity of laser vias was shown to be capable of supporting 112 Gb/s PAM-4 links (or equivalent signaling having 28 GHz (Nyquist) bandwidth). This original board design was then rebuilt using a different fabricator, and the test results revealed a significant decrease in the bandwidth of the vias.…
▽ More
An earlier study of a high layer-count test board using plated-through-hole (PTH) vias and a limited quantity of laser vias was shown to be capable of supporting 112 Gb/s PAM-4 links (or equivalent signaling having 28 GHz (Nyquist) bandwidth). This original board design was then rebuilt using a different fabricator, and the test results revealed a significant decrease in the bandwidth of the vias. These results led to the development of a set of design specifications that PCB vendors can easily validate, which will ensure that the use of high layer-count boards with PTH technology are viable for emerging 112 Gb/s PAM-4 links.
△ Less
Submitted 17 January, 2023;
originally announced April 2023.
-
The "AI+R"-tree: An Instance-optimized R-tree
Authors:
Abdullah-Al-Mamun,
Ch. Md. Rakin Haider,
Jianguo Wang,
Walid G. Aref
Abstract:
The emerging class of instance-optimized systems has shown potential to achieve high performance by specializing to a specific data and query workloads. Particularly, Machine Learning (ML) techniques have been applied successfully to build various instance-optimized components (e.g., learned indexes). This paper investigates to leverage ML techniques to enhance the performance of spatial indexes,…
▽ More
The emerging class of instance-optimized systems has shown potential to achieve high performance by specializing to a specific data and query workloads. Particularly, Machine Learning (ML) techniques have been applied successfully to build various instance-optimized components (e.g., learned indexes). This paper investigates to leverage ML techniques to enhance the performance of spatial indexes, particularly the R-tree, for a given data and query workloads. As the areas covered by the R-tree index nodes overlap in space, upon searching for a specific point in space, multiple paths from root to leaf may potentially be explored. In the worst case, the entire R-tree could be searched. In this paper, we define and use the overlap ratio to quantify the degree of extraneous leaf node accesses required by a range query. The goal is to enhance the query performance of a traditional R-tree for high-overlap range queries as they tend to incur long running-times. We introduce a new AI-tree that transforms the search operation of an R-tree into a multi-label classification task to exclude the extraneous leaf node accesses. Then, we augment a traditional R-tree to the AI-tree to form a hybrid "AI+R"-tree. The "AI+R"-tree can automatically differentiate between the high- and low-overlap queries using a learned model. Thus, the "AI+R"-tree processes high-overlap queries using the AI-tree, and the low-overlap queries using the R-tree. Experiments on real datasets demonstrate that the "AI+R"-tree can enhance the query performance over a traditional R-tree by up to 500%.
△ Less
Submitted 1 July, 2022;
originally announced July 2022.
-
Grid-SiPhyR: An end-to-end learning to optimize framework for combinatorial problems in power systems
Authors:
Rabab Haider,
Anuradha M. Annaswamy
Abstract:
Mixed integer problems are ubiquitous in decision making, from discrete device settings and design parameters, unit production, and on/off or yes/no decision in switches, routing, and social networks. Despite their prevalence, classical optimization approaches for combinatorial optimization remain prohibitively slow for fast and accurate decision making in dynamic and safety-critical environments…
▽ More
Mixed integer problems are ubiquitous in decision making, from discrete device settings and design parameters, unit production, and on/off or yes/no decision in switches, routing, and social networks. Despite their prevalence, classical optimization approaches for combinatorial optimization remain prohibitively slow for fast and accurate decision making in dynamic and safety-critical environments with hard constraints. To address this gap, we propose SiPhyR (pronounced: cipher), a physics-informed machine learning framework for end-to-end learning to optimize for combinatorial problems. SiPhyR employs a novel physics-informed rounding approach to tackle the challenge of combinatorial optimization within a differentiable framework that has certified satisfiability of safety-critical constraints. We demonstrate the effectiveness of SiPhyR on an emerging paradigm for clean energy systems: dynamic reconfiguration, where the topology of the electric grid and power flow are optimized so as to maintain a safe and reliable power grid in the presence of intermittent renewable generation. Offline training of the unsupervised framework on representative load and generation data makes dynamic decision making via the online application of Grid-SiPhyR computationally feasible.
△ Less
Submitted 24 May, 2023; v1 submitted 11 June, 2022;
originally announced June 2022.
-
Computational Resource Allocation for Edge Computing in Social Internet-of-Things
Authors:
Abdullah Khanfor,
Raby Hamadi,
Hakim Ghazzai,
Ye Yang,
Mohammad R. Haider,
Yehia Massoud
Abstract:
The heterogeneity of the Internet-of-things (IoT) network can be exploited as a dynamic computational resource environment for many devices lacking computational capabilities. A smart mechanism for allocating edge and mobile computers to match the need of devices requesting external computational resources is developed. In this paper, we employ the concept of Social IoT and machine learning to dow…
▽ More
The heterogeneity of the Internet-of-things (IoT) network can be exploited as a dynamic computational resource environment for many devices lacking computational capabilities. A smart mechanism for allocating edge and mobile computers to match the need of devices requesting external computational resources is developed. In this paper, we employ the concept of Social IoT and machine learning to downgrade the complexity of allocating appropriate edge computers. We propose a framework that detects different communities of devices in SIoT enclosing trustworthy peers having strong social relations. Afterwards, we train a machine learning algorithm, considering multiple computational and non-computational features of the requester as well as the edge computers, to predict the total time needed to process the required task by the potential candidates belonging to the same community of the requester. By applying it to a real-world data set, we observe that the proposed framework provides encouraging results for mobile computer allocation.
△ Less
Submitted 8 July, 2020;
originally announced July 2020.
-
Graph Neural Networks-based Clustering for Social Internet of Things
Authors:
Abdullah Khanfor,
Amal Nammouchi,
Hakim Ghazzai,
Ye Yang,
Mohammad R. Haider,
Yehia Massoud
Abstract:
In this paper, we propose a machine learning process for clustering large-scale social Internet-of-things (SIoT) devices into several groups of related devices sharing strong relations. To this end, we generate undirected weighted graphs based on the historical dataset of IoT devices and their social relations. Using the adjacency matrices of these graphs and the IoT devices' features, we embed th…
▽ More
In this paper, we propose a machine learning process for clustering large-scale social Internet-of-things (SIoT) devices into several groups of related devices sharing strong relations. To this end, we generate undirected weighted graphs based on the historical dataset of IoT devices and their social relations. Using the adjacency matrices of these graphs and the IoT devices' features, we embed the graphs' nodes using a Graph Neural Network (GNN) to obtain numerical vector representations of the IoT devices. The vector representation does not only reflect the characteristics of the device but also its relations with its peers. The obtained node embeddings are then fed to a conventional unsupervised learning algorithm to determine the clusters accordingly. We showcase the obtained IoT groups using two well-known clustering algorithms, specifically the K-means and the density-based algorithm for discovering clusters (DBSCAN). Finally, we compare the performances of the proposed GNN-based clustering approach in terms of coverage and modularity to those of the deterministic Louvain community detection algorithm applied solely on the graphs created from the different relations. It is shown that the framework achieves promising preliminary results in clustering large-scale IoT systems.
△ Less
Submitted 8 July, 2020;
originally announced July 2020.
-
Automated Service Discovery for Social Internet-of-Things Systems
Authors:
Abdullah Khanfor,
Hakim Ghazzai,
Ye Yang,
Mohammad Rafiqul Haider,
Yehia Massoud
Abstract:
In this paper, we propose to design an automated service discovery process to allow mobile crowdsourcing task requesters select a small set of devices out of a large-scale Internet-of-things (IoT) network to execute their tasks. To this end, we proceed by dividing the large-scale IoT network into several virtual communities whose members share strong social IoT relations. Two community detection a…
▽ More
In this paper, we propose to design an automated service discovery process to allow mobile crowdsourcing task requesters select a small set of devices out of a large-scale Internet-of-things (IoT) network to execute their tasks. To this end, we proceed by dividing the large-scale IoT network into several virtual communities whose members share strong social IoT relations. Two community detection algorithms, namely Louvain and order statistics local method (OSLOM) algorithms, are investigated and applied to a real-world IoT dataset to form non-overlap** and overlap** IoT devices groups. Afterwards, a natural language process (NLP)-based approach is executed to handle crowdsourcing textual requests and accordingly find the list of IoT devices capable of effectively accomplishing the tasks. This is performed by matching the NLP outputs, e.g., type of application, location, required trustworthiness level, with the different detected communities. The proposed approach effectively helps in automating and reducing the service discovery procedure and recruitment process for mobile crowdsourcing applications.
△ Less
Submitted 21 March, 2020;
originally announced March 2020.
-
Can We Predict the Scenic Beauty of Locations from Geo-tagged Flickr Images?
Authors:
Ch. Md. Rakin Haider,
Mohammed Eunus Ali
Abstract:
In this work, we propose a novel technique to determine the aesthetic score of a location from social metadata of Flickr photos. In particular, we built machine learning classifiers to predict the class of a location where each class corresponds to a set of locations having equal aesthetic rating. These models are trained on two empirically build datasets containing locations in two different citi…
▽ More
In this work, we propose a novel technique to determine the aesthetic score of a location from social metadata of Flickr photos. In particular, we built machine learning classifiers to predict the class of a location where each class corresponds to a set of locations having equal aesthetic rating. These models are trained on two empirically build datasets containing locations in two different cities (Rome and Paris) where aesthetic ratings of locations were gathered from TripAdvisor.com. In this work we exploit the idea that in a location with higher aesthetic rating, it is more likely for an user to capture a photo and other users are more likely to interact with that photo. Our models achieved as high as 79.48% accuracy (78.60% precision and 79.27% recall) on Rome dataset and 73.78% accuracy(75.62% precision and 78.07% recall) on Paris dataset. The proposed technique can facilitate urban planning, tour planning and recommending aesthetically pleasing paths.
△ Less
Submitted 24 March, 2018;
originally announced April 2018.
