Computer Science Faculty Research & Creative Works
Copyright (c) 2023 Missouri University of Science and Technology All rights reserved.
https://scholarsmine.mst.edu/comsci_facwork
Recent documents in Computer Science Faculty Research & Creative Works
enus
Wed, 07 Jun 2023 07:45:06 PDT
3600

Geodistributed Multitier Workload Migration Over Multitimescale Electricity Markets
https://scholarsmine.mst.edu/comsci_facwork/1298
https://scholarsmine.mst.edu/comsci_facwork/1298
Wed, 07 Jun 2023 06:37:15 PDT
Virtual machine (VM) migration enables cloud service providers (CSPs) to balance workload, perform zerodowntime maintenance, and reduce applications' power consumption and response time. Migrating a VM consumes energy at the source, destination, and backbone networks, i.e., intermediate routers and switches, especially in a Geodistributed setting. In this context, we propose a VM migration model called Low Energy Application Workload Migration (LEAWM) aimed at reducing the perbit migration cost in migrating VMs over Geodistributed clouds. With a Geodistributed cloud connected through multiple Internet Service Providers (ISPs), we develop an approach to find out the migration path across ISPs leading to the most feasible destination. For this, we use the variation in the electricity price at the ISPs to decide the migration paths. However, reduced power consumption at the expense of higher migration time is intolerable for realtime applications. As finding an optimal relocation is $\mathcal {NP}$Hard, we propose an Ant Colony Optimization (ACO) based biobjective optimization technique to strike a balance between migration delay and migration power. A thorough simulation analysis of the proposed approach shows that the proposed model can reduce the migration time by $25\%$–$30\%$ and electricity cost by approximately $25\%$ compared to the baseline.
]]>
Sourav Kanti Addya et al.

Sigmoid ActivationBased Long ShortTerm Memory for Time Series Data Classification
https://scholarsmine.mst.edu/comsci_facwork/1297
https://scholarsmine.mst.edu/comsci_facwork/1297
Wed, 07 Jun 2023 06:37:12 PDT
With the enhanced usage of Artificial Intelligence (AI) driven applications, the researchers often face challenges in improving the accuracy of the data classification models, while trading off the complexity. In this paper, we address the classification of time series data using the Long ShortTerm Memory (LSTM) network while focusing on the activation functions. While the existing activation functions such as sigmoid and tanh are used as LSTM internal activations, the customizability of these activations stays limited. This motivates us to propose a new family of activation functions, called logsigmoid, inside the LSTM cell for time series data classification, and analyze its properties. We also present the use of a linear transformation (e.g., log tanh) of the proposed logsigmoid activation as a replacement of the traditional tanh function in the LSTM cell. Both the cell activation as well as recurrent activation functions inside the LSTM cell are modified with logsigmoid activation family while tuning the log bases. Further, we report a comparative performance analysis of the LSTM model using the proposed and the stateoftheart activation functions on multiple public timeseries databases.
]]>
Sajal Das

Dynamic Mode Decomposition Type Algorithms for Modeling and Predicting Queue Lengths at Signalized Intersections with Short Lookback
https://scholarsmine.mst.edu/comsci_facwork/1296
https://scholarsmine.mst.edu/comsci_facwork/1296
Tue, 09 May 2023 13:11:10 PDT
This Article Explores a Novel DataDriven Approach based on Recent Developments in Koopman Operator Theory and Dynamic Mode Decomposition (DMD) for Modeling Signalized Intersections. on Signalized Intersections, Vehicular Flow and Queue Formation Have Complex Nonlinear Dynamics, Making System Identification, Modeling, and Controller Design Challenging. We Employ a DMDType Approach to Transform the Original Nonlinear Dynamics into Locally Linear InfiniteDimensional Dynamics. the DataDriven Approach Relies Entirely on SpatioTemporal Snapshots of the Traffic Data. We Investigate Several Key Aspects of the Approach and Provide Insights into the Usage of DMDType Algorithms for Application in Adaptive Signalized Intersections. to Validate the Obtained Linearized Dynamics, We Perform Prediction of the Queue Lengths at the Intersection and Compare the Results with the Benchmark Methods Such as ARIMA and Long Short Term Memory (LSTM). the Case Study Involves Intersection Pressure and Queue Lengths at Two Orlando Area Signalized Intersections during the Morning and Evening Peaks. It is Observed that DMDType Algorithms Are Able to Capture Complex Dynamics with a Linear Approximation to a Reasonable Extent. the Merits Include Faster Computation Times and Significantly Less Requirement for a "Lookback" (Training) Window.
]]>
Kazi Redwan Shabab et al.

Deep Meta QLearning based MultiTask Offloading in EdgeCloud Systems
https://scholarsmine.mst.edu/comsci_facwork/1295
https://scholarsmine.mst.edu/comsci_facwork/1295
Wed, 03 May 2023 11:11:06 PDT
ResourceConstrained Edge Devices Can Not Efficiently Handle the Explosive Growth of Mobile Data and the Increasing Computational Demand of ModernDay User Applications. Task Offloading Allows the Migration of Complex Tasks from User Devices to the Remote EdgeCloud Servers Thereby Reducing their Computational Burden and Energy Consumption While Also Improving the Efficiency of Task Processing. However, Obtaining the Optimal Offloading Strategy in a MultiTask Offloading DecisionMaking Process is an NPHard Problem. Existing Deep Learning Techniques with Slow Learning Rates and Weak Adaptability Are Not Suitable for Dynamic MultiUser Scenarios. in This Article, We Propose a Novel Deep MetaReinforcement LearningBased Approach to the MultiTask Offloading Problem using a Combination of FirstOrder MetaLearning and Deep QLearning Methods. We Establish the MetaGeneralization Bounds for the Proposed Algorithm and Demonstrate that It Can Reduce the Time and Energy Consumption of IoT Applications by Up to 15%. through Rigorous Simulations, We Show that Our Method Achieves NearOptimal Offloading Solutions While Also Being Able to Adapt to Dynamic EdgeCloud Environments.
]]>
Nelson Sharma et al.

<inlineFormula><texMath Notation="LaTeX">$\log$</texMath></inlineFormula>Sigmoid ActivationBased Long ShortTerm Memory for Time Series Data Classification
https://scholarsmine.mst.edu/comsci_facwork/1294
https://scholarsmine.mst.edu/comsci_facwork/1294
Wed, 03 May 2023 11:11:05 PDT
With the Enhanced Usage of Artificial Intelligence (AI) Driven Applications, the Researchers Often Face Challenges in Improving the Accuracy of the Data Classification Models, While Trading Off the Complexity. in This Paper, We Address the Classification of Time Series Data using the Long ShortTerm Memory (LSTM) Network While Focusing on the Activation Functions. While the Existing Activation Functions Such as Sigmoid and $\tanh$ Are Used as LSTM Internal Activations, the Customizability of These Activations Stays Limited. This Motivates Us to Propose a New Family of Activation Functions, Called $\log$Sigmoid, Inside the LSTM Cell for Time Series Data Classification, and Analyze its Properties. We Also Present the Use of a Linear Transformation (E.g., $\log \tanh$) of the Proposed $\log$Sigmoid Activation as a Replacement of the Traditional $\tanh$ Function in the LSTM Cell. Both the Cell Activation as Well as Recurrent Activation Functions Inside the LSTM Cell Are Modified with $\log$Sigmoid Activation Family While Tuning the $\log$ Bases. Further, We Report a Comparative Performance Analysis of the LSTM Model using the Proposed and the StateOfTheArt Activation Functions on Multiple Public TimeSeries Databases.
]]>
Priyesh Ranjan et al.

Programmable SoftwareDefined Testbed for Visible Light UAV Networks: Architecture Design and Implementation
https://scholarsmine.mst.edu/comsci_facwork/1293
https://scholarsmine.mst.edu/comsci_facwork/1293
Thu, 06 Apr 2023 06:24:32 PDT
As of Today, There Has Been Increasing Research on Designing Optimization Algorithms and Intelligent Network Control Methods for Visible Light Unmanned Aerial Vehicles (UAV) Networks to Provide Pervasive and Broadband Connections. for Those Theoretical Analysis based Algorithms, there is an Urgent Need to Have a Visible Light UAV Network Platform that Can Help Evaluate the Proposed Algorithms in RealWorld Scenarios. However, to the Best of Our Knowledge, there is Currently No Dedicated High Data Rate and Flexible Visible Light UAV Networking Prototype. to Bridge This Gap, in This Paper, We First Design a Novel Programmable SoftwareDefined Architecture for Visible Light UAV Networking, Including Control Plane, Network Plane, Signal Processing Chain and FrontEnds Plane, and Ground Facility Plane. We Then Implement a Prototype and Conduct Numerous Experiments to Validate the Feasibility of VisibleLight UAV Networks and Further Evaluate the System Performance Pertaining to Achievable Data Rate and Transmission Distance. the RealTime Video Streaming Experimental Results Show that Up to 550 Kbps Data Rate and a Maximum Distance of 7 Meters Can Be Achieved.
]]>
Yue Zhang et al.

No More Strided Convolutions or Pooling: A New CNN Building Block for LowResolution Images and Small Objects
https://scholarsmine.mst.edu/comsci_facwork/1292
https://scholarsmine.mst.edu/comsci_facwork/1292
Thu, 06 Apr 2023 06:24:31 PDT
Convolutional Neural Networks (CNNs) Have Made Resounding Success in Many Computer Vision Tasks Such as Image Classification and Object Detection. However, their Performance Degrades Rapidly on Tougher Tasks Where Images Are of Low Resolution or Objects Are Small. in This Paper, We Point Out that This Roots in a Defective Yet Common Design in Existing CNN Architectures, Namely the Use of Strided Convolution And/or Pooling Layers, Which Results in a Loss of FineGrained Information and Learning of Less Effective Feature Representations. to This End, We Propose a New CNN Building Block Called SPDConv in Place of Each Strided Convolution Layer and Each Pooling Layer (Thus Eliminates Them Altogether). SPDConv is Comprised of a SpaceToDepth (SPD) Layer Followed by a NonStrided Convolution (Conv) Layer, and Can Be Applied in Most If Not All CNN Architectures. We Explain This New Design under Two Most Representative Computer Vision Tasks: Object Detection and Image Classification. We Then Create New CNN Architectures by Applying SPDConv to YOLOv5 and ResNet, and Empirically Show that Our Approach Significantly Outperforms StateOfTheArt Deep Learning Models, Especially on Tougher Tasks with LowResolution Images and Small Objects. We Have OpenSourced Our Code at Https://github.com/LabSAINT/SPDConv.
]]>
Raja Sunkara et al.

Classical Quotient Rings of Group Rings
https://scholarsmine.mst.edu/comsci_facwork/1290
https://scholarsmine.mst.edu/comsci_facwork/1290
Fri, 31 Mar 2023 11:39:51 PDT
Throughout G Will Denote a Free Abelian Group and Z(R) the Right Singular Ideal of a Ring R. a Ring R is a ClRing If R is (Goldie) Right Finite Dimensional, R/Z(R) is Semiprime, Z(R) is Rationally Closed, and Z(R) Contains No Closed Uniform Right Ideals. We Prove that R is a ClRing If and Only If the Group Ring RG is a C1Ring. If RG Has the Additional Property that BRG is Dense Whenever B is a Right NonzeroDivisor, Then the Complete Ring of Quotients of RG is a Classical Ring of Quotients. © 1976 Taylor and Francis Group, LLC.
]]>
R. (Ralph) W. Wilkerson

Finite Dimensional Group Rings
https://scholarsmine.mst.edu/comsci_facwork/1291
https://scholarsmine.mst.edu/comsci_facwork/1291
Fri, 31 Mar 2023 11:39:51 PDT
A Ring is Right Finite Dimensional If It Contains No Infinite Direct Sum of Right Ideals. We Prove that If a Group G is Finite, Free Abelian, or Finitely Generated Abelian, then a Ring R is Right Finite Dimensional If and Only If the Group Ring RG is Right Finite Dimensional. a Ring R is a SelfInjective Cogenerator Ring If Rn is Injective and RR is a Cogenerator in the Category of Unital Right /{Modules; This Means that Each Right Unital AModule Can Be Embedded in a Direct Product of Copies of R. Let G Be a Finite Group Where the Order of G is a Unit in R. Then the Group Ring RG is a Selfinjective Cogenerator Ring If and Only If R is a SelfInjective Cogenerator Ring. Additional Applications Are Given. © 1973 American Mathematical Society.
]]>
Ralph W. Wilkerson

The Double Eigenvalue Problem; Including Numerical Solutions
https://scholarsmine.mst.edu/comsci_facwork/1289
https://scholarsmine.mst.edu/comsci_facwork/1289
Fri, 31 Mar 2023 11:39:50 PDT
John Gregory et al.

An Approximation Theory for Conjugate Surfaces and Solutions of Elliptic Multiple Integral Problems: Application to Numerical Solutions of Generalized Laplace's Equation
https://scholarsmine.mst.edu/comsci_facwork/1287
https://scholarsmine.mst.edu/comsci_facwork/1287
Fri, 31 Mar 2023 11:39:49 PDT
An Approximation Theory is Given for a Class of Elliptic Quadratic Forms Which Include the Study of Conjugate Surfaces for Elliptic Multiple Integral Problems. These Ideas Follow from the Quadratic Form Theory of Hestenes, Applied to Multiple Integral Problems by Dennemeyer, and Extended with Applications for Approximation Problems by Gregory. the Application of This Theory to a Variety of Approximation Problem Areas in This Setting is Given. These Include Conjugate Surfaces and Conjugate Solutions in the Calculus of Variations, Oscillation Problems for Elliptic Partial Differential Equations, Eigenvalue Problems for Compact Operators, Numerical Approximation Problems, And, Finally, the Intersection of These Problem Areas. in the Final Part of This Paper the Ideas Are Specifically Applied to the Construction and Counting of Negative Vectors in Order to Obtain New Numerical Methods for Solving LaplaceType Equations and to Obtain the "EulerLagrange Equations" for SymmetricBanded Tridiagonal Matrices. in This New Result (Which Will Allow the Reexamination of Both the Theory and Applications of Symmetric banded Matrices) One Can Construct, in a Meaningful Way, Negative Vectors, Oscillation Vectors, Eigenvectors, and Extremal Solutions of Classical Problems as Well as Efficient Algorithms for the Numerical Solution of Partial Differential Equations. Numerical Examples (Test Runs) Are Given. © 1982.
]]>
John Gregory et al.

New Numerical Algorithms for Eigenvalues and Eigenvectors of Second Order Differential Equations
https://scholarsmine.mst.edu/comsci_facwork/1288
https://scholarsmine.mst.edu/comsci_facwork/1288
Fri, 31 Mar 2023 11:39:49 PDT
In This Paper We Present for the First Time an Accurate, Fast, and Easy to Implement Numerical Algorithm to Find the Eigenvalues and Eigenvectors of the Equation L(X;λ) = (Rx)+px+λqx = 0. These Ideas Follow from a Theory of Quadratic Forms Given by the First Author and Will Be Applicable in a Wide Variety of Eigenvalue Problems. We Include Test Runs to Demonstrate that the Accuracy of Our Methods Are Superior to More Conventional Projection Methods. © 1981, Taylor & Francis Group, LLC. All Rights Reserved.
]]>
John Gregory et al.

Linear Maps of Holomorphic Functions on Commutative Algebras
https://scholarsmine.mst.edu/comsci_facwork/1286
https://scholarsmine.mst.edu/comsci_facwork/1286
Fri, 31 Mar 2023 11:39:48 PDT
Generalized Function Theory on Finite Dimensional Linear Associative Algebras over the Field of Reals Has Produced Many Forms of Generalized Regularity. of Particular Interest Are the Forms of Regularity Known as SRegularlty and FRegularity. It is Well Known that Functions Regular in One Sense Need Not Be Regular in the Other. Here We Prove a Transformation Theorem Which Allows the Formation of FRegular Functions from ․Regular Functions. Furthermore, a Classification Theorem is Proved that Determines the Algebras for Which This Transformation Property Holds. © 1982 by Marcel Dekker, Inc.
]]>
Ralph W. Wilkerson

PROLOG.
https://scholarsmine.mst.edu/comsci_facwork/1284
https://scholarsmine.mst.edu/comsci_facwork/1284
Fri, 31 Mar 2023 11:39:47 PDT
A Description is Given of Prolog, a Contraction of Programming in Logic, Which Uses the Formalism of Mathematical Logic as its Primary Design Principle. the Structure of Prolog is Examined, and a Database Program is Described to Illustrate its application. an Application to an Artificial Intelligence Problem, the Towers of Hanoi, is Also Given.
]]>
Ralph W. Wilkerson

Symbolic Computation and the Dirichlet Problem
https://scholarsmine.mst.edu/comsci_facwork/1285
https://scholarsmine.mst.edu/comsci_facwork/1285
Fri, 31 Mar 2023 11:39:47 PDT
Ralph W. Wilkerson

A Logic Programming Model of the Game of Sprouts
https://scholarsmine.mst.edu/comsci_facwork/1283
https://scholarsmine.mst.edu/comsci_facwork/1283
Fri, 31 Mar 2023 11:39:46 PDT
The Game of Sprouts Has Intrigued Mathematicians for Nearly Twenty Years. This Paper Describes a Representation Scheme Which Simplifies Much of the Geometry of the Game. using This Representation, We Develop a Prolog Program Which Will Play Sprouts. It is Hoped that the Program Will Prove to Be a Useful Research Tool in Finding the Key to a Winning Strategy for Sprouts and that the Representation Will Serve as a Useful Model for Studying Planar Graphs. © 1987, ACM. All Rights Reserved.
]]>
Ralph M. Butler et al.

A Logic Programming Model of the Game of Sprouts
https://scholarsmine.mst.edu/comsci_facwork/1282
https://scholarsmine.mst.edu/comsci_facwork/1282
Fri, 31 Mar 2023 11:39:45 PDT
The Game of Sprouts Has Intrigued Mathematicians for Nearly Twenty Years. This Paper Describes a Representation Scheme Which Simplifies Much of the Geometry of the Game. using This Representation, We Develop a Prolog Program Which Will Play Sprouts. It is Hoped that the Program Will Prove to Be a Useful Research Tool in Finding the Key to a Winning Strategy for Sprouts and that the Representation Will Serve as a Useful Model for Studying Planar Graphs.
]]>
Ralph M. Butler et al.

Automated Circuit Diagnosis using First Order Logic Tools
https://scholarsmine.mst.edu/comsci_facwork/1280
https://scholarsmine.mst.edu/comsci_facwork/1280
Fri, 31 Mar 2023 11:39:44 PDT
While Numerous Diagnostic Expert Systems Have Been Successfully Developed in Recent Years, They Are Almost Uniformly based on Heuristic Reasoning Techniques (I.e., Shallow Knowledge) in the Form of Rules. This Paper Reports on an Automated Circuit Diagnostic Tool based on Reiter's Theory of Diagnosis. in Particular, this is a Theory of Diagnosis based on Deep Knowledge (I.e., Knowledge based on Certain Design Information) and using First Order Logic as the Representation Language. the Inference Mechanism Which is Incorporated as Part of the Diagnostic Tool is a Refutation based Theorem Prover using Rewriting Systems for Boolean Algebra Developed by Hsiang. Consequently, the Diagnostic Reasoning Tool is Broadly based on Reiter's Model but Incorporates Complete Sets of Reductions for Boolean Algebra to Reason over EquaTional Descriptions of the Circuits to Be Analyzed. the Refutational Theorem Prover Uses an Associative Commutative Identity Unification Algorithm Described by Hsiang but Requires Additional Focusing Techniques in Order to Be Appropriate for Diagnosing Circuits. a Prototype Version of the Mainline Diagnostic Program Has Been Developed and Has Been Successfully Demonstrated on Several Small but Nontrivial Combinational Circuit Examples.
]]>
Barbara Smith et al.

A Routing Algorithm for Three Stage Rearrangeable Clos Networks
https://scholarsmine.mst.edu/comsci_facwork/1281
https://scholarsmine.mst.edu/comsci_facwork/1281
Fri, 31 Mar 2023 11:39:44 PDT
Ralph W. Wilkerson

A Correction to the Algorithm in Reiter's Theory of Diagnosis
https://scholarsmine.mst.edu/comsci_facwork/1279
https://scholarsmine.mst.edu/comsci_facwork/1279
Fri, 31 Mar 2023 11:39:43 PDT
Reiter [3] Has Developed a General Theory of Diagnosis based on First Principles. His Algorithm Computes All Diagnoses Which Explain the Differences between the Predicted and Observed Behavior of a Given System. Unfortunately, Reiter's Description of the Algorithm is Incorrect in that Some Diagnoses Can Be Missed under Certain Conditions. This Note Presents a Revised Algorithm and a Proof of its Correctness. © 1989.
]]>
Russell Greiner et al.