PHD, Kent State University, 2021, College of Arts and Sciences / Department of Computer Science

A major challenge in modern multiprocessor computer programming is concurrency control: (i) how to coordinate accesses to memory locations shared among concurrently executing tasks and (ii) how to ensure that the computation is correct. The traditional approach is to use mutexes and locks but they have many drawbacks such as deadlocks. This dissertation explores the recently emerged paradigm of transactional memory for multiprocessor programming. In transactional memory, program code is split into transactions, blocks of code containing a sequence of read and write operations that appear to execute atomically to a set of shared resources. When transactions access the same shared resources at the same time, conflict occurs, and transactions may need to abort. A transaction commits if either no conflict occurs, or conflicts are resolved. Conflicts are typically resolved through a transaction scheduling algorithm. This dissertation explores transactional memory in the context of shared memory multiprocessor systems where concurrent tasks interact through reading and writing the same main memory as well as distributed multiprocessor systems where concurrent tasks interact by sending messages to each other. The main difference between shared memory and distributed systems is that there is non-uniformity in memory access latency in distributed systems, which is not the case in shared memory systems. This non-uniformity is vital and affects not only the total execution time of all concurrent tasks but also other related network parameters such as communication cost and congestion. For uniform latency, the focus is mainly on minimizing total execution time of all concurrent tasks. This dissertation develops several novel results in both shared memory and distributed multiprocessor systems. In shared memory systems, it develops a versioning method and a transaction scheduling mechanism. Specifically, it proposes an adaptive versioning method that switches between eage (open full item for complete abstract)

Committee: Gokarna Sharma (Advisor); Feodor F. Dragan (Committee Member); Mikhail Nesterenko (Committee Member); Murali Shanker (Committee Member) Subjects: Computer Science

Doctor of Philosophy, The Ohio State University, 2016, Electrical and Computer Engineering

Electric power grid is experiencing a major paradigm shift toward a more reliable, efficient, and environmentally friendly grid. The concept of microgrid is introduced to integrate distributed renewable generation in proximity to demands for both environmental and power-efficient promises. A microgrid can be disconnected, or "islanded", from the main grid and operates on its own, providing energy to remote areas or during faults of the main grid for better reliability. Islanded microgrids inherit several different properties from traditional power grids, including uncertain and limited generation, mixed R/X ratio lines, and lack of power inertia from synchronous generators. Those properties pose new challenges for the stable operation of islanded microgrids. The dissertation is dedicated to addressing the control challenges of islanded microgrids. The contribution is twofolds. First, we propose a polynomial time optimal power flow (OPF) solver which finds an optimal operating point for the inverters of the distributed energy resources. The proposed algorithm can account for the cost functions on the reactive generation that are common in microgrids. It also brings new understanding on the conjectures of exact semidefinite programming (SDP) convex relaxation on the OPF problem. Furthermore, we show that without the load over-satisfaction assumption usually seen in the literature, a near global optimum can be found for the OPF problem with arbitrary convex quadratic cost functions. The results are important to both microgrids and the classical OPF problem. Our second major contribution is developing a novel distributed controller that addresses the control challenges originated from limited generation, mixed R/X ratio lines, and lack of power inertia properties of islanded microgrids. The proposed controller can ensure proportional active and reactive power sharing and frequency synchronization while respecting the voltage constraints. Variances of the distributed (open full item for complete abstract)

Committee: Wei Zhang (Advisor); Kevin Passino (Committee Member); Andrea Serrani (Committee Member); Krishnaswamy Srinivasan (Other) Subjects: Electrical Engineering; Mechanical Engineering

MS, University of Cincinnati, 2006, Engineering : Computer Engineering

A majority of group communication applications in cellular-based heterogeneous wireless setups entail secure data exchange. The problem can be effectively tackled if the underlying cellular infrastructure is used to provide an authentication backbone to the security associations. We propose a novel distributed ID based key exchange mechanism using shared polynomials in which the shares are generated by the communicating groups. Our idea employs a mechanism where the Base Stations (BSs) carry out an initial key generation by a polynomial in a distributed manner and then pass on the key material to the Mobile Stations (MSs). The multi-interface MSs can now securely communicate over interfaces other than cellular. The scheme incorporates symmetric polynomials, which are chosen by the BS acting as polynomial distributors. Simulations done to measure performance have shown encouraging results.

Committee: Dr. Agrawal Dharma (Advisor) Subjects: Computer Science

PhD, University of Cincinnati, 2004, Engineering : Mechanical Engineering

The dissertation is to address the need, in contact mechanics, of efficient and effective solutions to certain 3-D contact problems. The solutions developed here are based on underlying analytical solutions to pyramidal loading elements. This feature, along with other characteristics, distinguishes this method from other numerical solutions. The research work is logically divided into three subsequent parts, each of which addresses a particular aspect of the project: (1) Developed analytical solution sets in closed form to pyramidal loading profiles. First, a set of Boussinesq-Curruti equations to linear/bilinear distribution of normal and tangential loading over a triangular area are derived and evaluated. Second, solution sets to normal and tangential surface loading pyramids are constructed. The work provides a solution set to a basic loading element, which is the foundation of the development of effective and efficient semi-analytical solutions to 3-D contact problems with general geometry and loading profile. (2) Developed a semi-analytical approach (non-incremental algorithm) to 3-D normal contact problems with friction. This approach treats normal contact (indentation) phenomenon as a static problem. Based on fully coupled governing equations, the algorithm of contact detecting and stick/slip partitioning is designed as nested iterations, to fulfill contact boundary conditions. The computation shows that it is an efficient algorithm. Numerical examples are presented to show the accuracy and efficiency of the method.(3) Developed a semi-analytical approach (incremental algorithm) to 3-D contact problems with friction. This approach treats contact as a dynamic problem. The general dynamic models are simplified into quasi-static models in many practical cases that inertial force can be ignored. The incremental algorithm is designed to solve the quasi-static problems. The computation shows that the algorithm works very well for cases featuring both similar and di (open full item for complete abstract)

Committee: Dr. EDWARD BERGER (Advisor) Subjects: Engineering, Mechanical

MS, University of Cincinnati, 2003, Engineering : Computer Science

Wireless Personal Area Network (WPAN) is playing an important role in providing networking infrastructure for short-range devices. Bluetooth is a low-power, low-cost, short-range wireless communication system, thus constituting a typical example of WPAN. In a Bluetooth ad hoc network, up to eight Bluetooth devices can communicate with each other in a special network called piconet. Each piconet has one master and up to seven active slaves. Scatternet can be formed by interconnecting several piconets, the shared device between two piconets being known as a bridge. The master devices and the bridge devices in the scatternet invariably have more responsibilities. We focus on optimum role assignment in the scatternet so that resource-constrained devices do not assume the role of a master and/or a bridge. In this dissertation, we present a distributed resource aware scalable scatternet formation algorithm. Though the proposed algorithm has two phases, it is not necessary that all the nodes be in the same phase at a given instant of time. The formed scatternet has the following properties: 1. All the devices may not be necessarily in each other's transmission range. 2. Any device will be a member of at the most two piconets. 3. Higher resource-weighted devices are preferred over lower resource-weighted devices to perform the role of a master and/or a bridge. 4. Slave/Slave bridges are preferred over Master/Slave bridges. We have done extensive simulations that show that the number of messages transmitted by low resource-weighted devices to be significantly lesser than the number of messages exchanged by high resource-weighted devices, and the number of master/slave bridges is remarkably less as compared to the number of slave/slave bridges. Proposed scatternet formation scheme being a self-organizing algorithm, finds its applications in wireless sensor networks. We use the backbone formed by the self-organizing algorithm as a collaborative framework for event detection in (open full item for complete abstract)

Committee: Dr. Dharma P. Agrawal (Advisor) Subjects: Computer Science

PhD, University of Cincinnati, 2003, Engineering : Computer Science and Engineering

Most computer algorithms have been designed for situations in which all relevant data is stored at a single Computer site. This is the classical model of a computer based information and control system. The emerging networked knowledge environment requires a significant move away from this classical model. In these situations of geographically distributed but networked systems, the data relevant for a computation may exist in a number of different databases residing at different network sites. An efficient system for computations with such distributed data would work by doing as much work at local sitesas possible and then communicating minimum required information among the sites. This is much more efficient than transferring the complete databases to a single site, Join these databases, and then execute algorithms with this data. A common constraint in these situations is that the databases cannot be moved to other network sites due to data-security, size, privacy or data-ownership considerations. Also, for some huge databases it may not be feasible to store and compute with them at one computer site. A number of partitions of this database may be stored at different sites and a set of cooperative algorithms run across the network that produce exactly the same results that would have been obtained if the database were processed at some single site. In this dissertation we present the results of development, validation, implementation and complexity analysis of the decomposable versions of a number of algorithms. Specifically, decomposable algorithms for following tasks have been investigated: finding a path between two vertices , computing the shortest paths spanning tree and the minimum spanning tree for a graph stored as components across various sites of a networks, non-hierarchical clustering in horizontally and vertically partitioned datasets across a number of geographically distributed databases. The main objective of our algorithms is their self-decomposab (open full item for complete abstract)

Committee: Dr. Raj Bhatnagar (Advisor) Subjects: Computer Science

MS, University of Cincinnati, 2002, Engineering : Computer Science

The pair of nearest neighbors in a given set S of n data points in d-dimensional space in distributed databases can be found by Centralized and Decomposable methods. In the Centralized method, all the data points are brought to one computer and then computations are done to find to the pair of nearest neighbors. A new decomposable algorithm has been developed to find a pair of nearest neighbors in a given set S of n data points in d-dimensional space in distributed databases. In this algorithm, individual computers also perform significant amount of computation as compared to the brute force or Centralized method. This results in less transmission of data. Results based on the proposed algorithm are found to give same result as the brute force method.

Committee: Dr. Raj Bhatnagar (Advisor) Subjects: Computer Science

MS, University of Cincinnati, 2001, Engineering : Computer Engineering

Optimal scheduling of parallel tasks with some precedence relationship, onto a distributed-memory machine is known to be a strong NP-hard problem. The complexity of the problem increases when task scheduling is to be done on a network of heterogeneous workstations (NoWs), where workstations may not be identical and may take different amount of time to execute the same task. This dissertation presents a Task duplication based scheduling Algorithm for Network of Heterogeneous workstations (TANH), with complexity O(V 2 ), which provides optimal results for applications represented by Directed Acyclic Graphs (DAGs), provided some simple conditions on task computation and network communication time could be satisfied. The performance of the algorithm is illustrated by comparing the scheduling time with an existing scheme BIL, for heterogeneous systems. We also observe that TANH provides speed-ups of 6 to 40 for some practical DAGs with upto 3000 nodes and the number of edges varying from 4000 to as high as 25000 edges, if adequate duplication is allowed. The flexibility of scaling to higher or lower number of workstations, as per their availability is also discussed. Some problems for future work are also briefly outlined.

Committee: Dr. Dharma P. Agrawal (Advisor) Subjects:

Master of Science, The Ohio State University, 2010, Electrical and Computer Engineering

In this research, we investigate designing a smart lighting system. By extending and enhancing the centralized and distributed control algorithms we try to address the lighting control problem and design a robust smart lighting system. The purpose of implementing various control strategies is to come up with a strategy that optimizes the power consumed by the system and its robustness to the problems of cross-illumination, external light disturbances, delays in the communication network, and the network topologies used for communication. The functionalities we try to achieve are uniform lighting, user-defined preference based lighting, maximum lighting mode and energy savings lighting mode by utilizing daylight. We study the performance of each control strategy and present a comparative analysis between the best strategies.

Committee: Kevin Passino PhD (Advisor); Vadim Utkin PhD (Committee Member) Subjects: Electrical Engineering; Energy

Doctor of Philosophy, The Ohio State University, 2005, Computer and Information Science

Ad Hoc wireless networks have been gaining importance in the communication world for the past decade. Wireless network extends the access to the network by removing the restriction of physical wires. Ad hoc network further improves the network coverage and availability to places without infrastructure support. Clock synchronization and dominating set are two of the fundamental issues in the wireless ad hoc networks. They are important for the correctness and/or performance of many protocols and applications. We focus on IEEE 802.11 wireless ad hoc networks in this dissertation due to the wide deployments of 802.11 networks. The theories and practices are definitely extensible to other types of ad hoc networks. IEEE 802.11 wireless network depends heavily on the distribution of timing information to all the stations in the network. Clock synchronization is important for frequency hopping, power saving and wireless medium reservation. We review the Timing Synchronization Function (TSF) of Ad Hoc mode defined in the 802.11 standard. It is well-known that the 802.11 TSF is not scalable. We carefully analyze the root causes of the scalability problem and design new schemes to overcome the problem. Our new schemes show great improvement over the 802.11 TSF and other solutions in the fields. Our solutions have nice characteristics: scalable, accurate, bounded and adaptive to station mobility. We are able to control the maximum clock differences under 25 micro seconds and 50 micro seconds in single hop and multihop networks respectively. The performance improvement is at least 200% or more compared with the current solutions. Dominating set has been widely used in multihop ad hoc networks (MANET) by numerous routing, broadcast and collision avoidance protocols. The problem to construct a minimum sized dominating set is known to be NP-hard. We propose a protocol that is simple, distributed, inexpensive, and adaptive to station mobility. We show that our protocol can construc (open full item for complete abstract)

Committee: Ten-Hwang Lai (Advisor) Subjects: Computer Science

Master of Science, The Ohio State University, 2013, Electrical and Computer Engineering

This research is motivated by the world's fast-growing demand for energy savings. Due to the increasing cost of fossil fuels (e.g., oil, coal, natural gas), research has been conducted to effectively reduce the electricity consumption in office buildings by means of employing smart lighting. This thesis investigates the implementation of an adaptive and nonadaptive fuzzy control for a smart light experimental testbed. The objective is to accurately regulate the light level across the experimental testbed to a desired voltage reference value, and to test the performance of the fuzzy controllers under cross-illumination effects, and bulb and sensor failures. As an initial approach, a decentralized (i.e., no communication between controllers) nonadaptive fuzzy controller is implemented and applied to the experimental testbed. This approach is convenient for this type of experimental testbed where a mathematical model of the plant is not available and heuristic information about how to control the system is sufficient. The nonadaptive fuzzy controller, when properly tuned, is able to achieve uniform lighting across the entire testbed floor in most of the tested situations but it fails whenever an on/off light bulb failure is introduced. In order to attain uniform lighting for complex failures, a fuzzy model reference learning controller (i.e., adaptive fuzzy) is developed for the experimental testbed, and this algorithm proves to be able to adapt to uncertainties such as disturbances and failures via a learning mechanism.

Committee: Kevin M. Passino Prof. (Advisor); Wei Zhang Prof. (Committee Member) Subjects: Electrical Engineering