Digital Commons Network^™

Automated Manpower Rostering: Techniques And Experience, C. M. Khoong, Hoong Chuin Lau, L. W. Chew

Research Collection School Of Computing and Information Systems

We present ROMAN, a comprehensive, generic manpower rostering toolkit that successfully handles a wide spectrum of work policies found in service organizations. We review the use of various techniques and methodologies in the toolkit that contribute to its robustness and efficiency, and relate experience gained in addressing manpower rostering problems in industry.

Analysis Of Myoelectrical Signals For Building A Dextrous Hand, Christopher T. Creel, Kishan Mehrotra, Chilukuri K. Mohan, Sanjay Ranka

Electrical Engineering and Computer Science - Technical Reports

We analyze techniques for myoelectrical signals classification for the purpose of designing a multifunctional prosthetic device for human amputees. The main advantage of our system over existing models is that it is more robust, easier to work with, more general, and efficient enough to run in real time. We achieve this with the help of "Supervised Growing Cell Structures." an artificial neural network model designed by Fritzke [10]. The current paper focuses on the flexion of the index, middle and ring fingers, as these are the most difficult movements to tackle.

Hardware Assists For High Performance Computing Using A Mathematics Of Arrays, H. Pottinger, W. Eatherton, L. Mullin, T. Schiefelbein

Computer Science Technical Reports

Work in progress at the University of Missouri - Rolla on hardware assists for high performance computing and communication is presented. This research consists of a field programmable gate array based rapid prototyping environment (Chameleon) being used to evaluate hardware architectures for speedup of array partitioning and routing algorithms. These algorithms have been developed using a Mathematics of Arrays (MOA) and are optimal in the sense that they reduce normal array indexing operations to a series of primitive additions based on array shape. The method is simple yet powerful, produces algorithms which are provably correct, and are independent of dimensionality. …

An Fpga Based Reconfigurable Coprocessor Board Utilizing A Mathematics Of Arrays, H. Pottinger, W. Eatherton, T. Schiefelbein, L. Mullin, R. Ziegler

Computer Science Technical Reports

Work in progress at the University of Missouri-Rolla on hardware assists for high performance computing is presented. This research consists of a novel field programmable gate array (FPGA) based reconfigurable coprocessor board (the Chameleon Coprocessor) being used to evaluate hardware architectures for speedup of array computation algorithms. These algorithms have been developed using a Mathematics of Arrays (MOA) and are optimal in the sense that they reduce normal array cartesian indexing operations to a series of primitive additions and offsets based on array shape. The method is simple yet powerful, produces algorithms which are provably correct, and are independent of …

A Recursive Least Squares Training Algorithm For Multilayer Recurrent Neural Networks, Q. Xu, K. Krishnamurthy, Bruce M. Mcmillin, Wen Feng Lu

Mechanical and Aerospace Engineering Faculty Research & Creative Works

Recurrent neural networks have the potential to perform significantly better than the commonly used feedforward neural networks due to their dynamical nature. However, they have received less attention because training algorithms/architectures have not been well developed. In this study, a recursive least squares algorithm to train recurrent neural networks with an arbitrary number of hidden layers is developed. The training algorithm is developed as an extension of the standard recursive estimation problem. Simulated results obtained for identification of the dynamics of a nonlinear dynamical system show promising results.

Qualitative Study Of The Symplectic Stormer-Verlet Integrator, D. J. Hardy, D. I. Okunbor

Computer Science Technical Reports

Symplectic numerical integrators, such as the Stormer-Verlet method, are useful in preserving certain important properties that are not preserved by conventional numerical integrators. This paper analyzes the Stormer-Verlet method as applied to the simple harmonic model, which is an important model for molecular dynamics simulations. Restricting our attention to the one-dimensional case, both the exact solution and the Stormer-Verlet solution to this model are expressed as functions of the number of time steps taken, and then both of these functions are interpreted geometrically. The paper shows the existence of an upper bound on the error from the Stormer-Verlet method, and …

Indexing And Distributing A General Patterned Spare Array, Daria Dooling, Lenore Mullin

Computer Science Technical Reports

High performance computing and communication (HPCC) requires that space and time usage are kept to a minimum when solving very large scientific problems. These problems often are represented by sparse systems, and in particular banded systems. In this paper, a new banded array index algorithm is presented that uses substantially less memory than the BLAS (Basic Linear Algebra Subroutines) representation. The algorithm presented here works for multi-dimensional arrays and on all bandwidths. It is visualized as a layer between the applications and the data, thus shielding the caller from the complexities of the sparse representation. Information is returned to the …

Ccsp - A Formal System For Distributed Program Debugging, Beth Arrowsmith, Bruce Mcmillin

Computer Science Technical Reports

One of the major problems with programming in a parallel/distributed environment is the difficulty in debugging the programs due to their complex interactions. One can have results that appear random, but when given a complete knowledge of the specific run-time behavior, the results are foregone. Unfortunately, this complete knowledge is not generally attainable in a distributed system. In order to develop a system for debugging distributed program and, for the more general case, ensuring their correctness at run-time, we built a distributed execution enviromnent based on Hoare's CSP [5] which allows for the execution and evaluation of embedded assertions within …

Parallel Adaptive Mesh Refinement Algorithms, Mustafa Keskin, Fikret Ercal

Computer Science Technical Reports

This study is aimed at developing 2D parallel adaptive mesh refinement algorithms for engineering applications such as penetration mechanics, manufacturing and combustion which use the finite element method for simulation. A new algorithmic approach, called piecewise adaptive mesh refinement for parallelization, is developed and implemented to run on a sequential machine. Performance of the new method is compared against a conventional advancing front technique proposed earlier. A parallel implementation model to run on a massively parallel computer is proposed and its expected efficiency is analyzed theoretically. Furthermore, some issues concerning parallelization and task partitioning are also discussed. In addition, a …

A Deterministic Membership Algorithm In Asynchronous Distributed Systems, P. Su, B. Mcmillin, A. Dekock

Computer Science Technical Reports

This work presents a deterministic asynchronous membership protocol (AMP). Conceptually a membership algorithm provides a means of recruiting members (machines) in a distributed system. Currently most of the research in membership recruiting emphasizes how to obtain the members after a machine fail-stop has occurred during the system operation, and the membership algorithms proposed arc non-deterministic. This work extends the research to develop a deterministic membership algorithm to recruit machines dynamically at the system cold-start. Once it is possible to recruit machines at system cold-start, then it is trivial to recruit members after a machine fail-stop has occurred during the system …

Using Temporal Subsumption To Generate Efficient Error-Detecting Distributed Algorithms, Martina Schollmeyer, Bruce Mcmillin

Computer Science Technical Reports

Distributed algorithms can use executable assertions, which can be derived from program verificatioll or specified in an ad hoc manner, to detect. errors at run-time. However, there may be more assertions available than are really unecessary, and embedding all of them into the program to be checked at run-time would make error-detection very inefficient. For safety-critical systems we often need to make decisions very quickly and cannot allow too much time to be spent on error detection.

The new technique of temporal submission, which is introduced in this dissertation, examines the dependencies between the individual assertions along program execution paths. …

The Formal Description Of Resource Deadlock In Distributed Systems, P. Lu, B. Mcmillin

Computer Science Technical Reports

Deadlock detection is a fundamental problem in a distributed system and has been extensively studied in the past few years. Many distributed deadlock detection/resolution alg·orithms have been proposed, but most of them either have not given a correctness proof or have given an informal proof by using intuitive operational arguments. Informal arguments are prone to errors and many of the published algorithms have been fouud to be incorrect. In trying to avoid this situation, the development of a formal approach to the algorithm correctness proof is required.

In this work, a formal resource deadlock model with global and local clocks …

Production And Decay Of D1 (2420)0 And D2* (2460)0, Avery, P.; Et Al., M. Thulasidas

Research Collection School Of Computing and Information Systems

No abstract provided.

Identification Of Cutting Force In End Milling Operations Using Recurrent Neural Networks, Q. Xu, K. Krishnamurthy, Bruce M. Mcmillin, Wen Feng Lu

Mechanical and Aerospace Engineering Faculty Research & Creative Works

The problem of identifying the cutting force in end milling operations is considered in this study. Recurrent neural networks are used here and are trained using a recursive least squares training algorithm. Training results for data obtained from a SAJO 3-axis vertical milling machine for steady slot cuts are presented. The results show that a recurrent neural network can learn the functional relationship between the feed rate and steady-state average resultant cutting force very well. Furthermore, results for the Mackey-Glass time series prediction problem are presented to illustrate the faster learning capability of the neural network scheme presented here

Extracting Parallelism At Compile-Time Through Dependence Analysis & Cloning Techniques In An Object-Based Paradigm, Binoy Ravindran

Theses

The construct of Abstract Data Type (ADT) modules and Abstract Data Object (ADO) modules supported by most object-based languages are a great source for developing reusable code. To improve the run time performance of such object-based programs, we consider the asynchronous remote procedure call (ARPC) model of parallel execution, in which concurrency is achieved by having the caller and the callee (which are module instances) running on different processors. Frequently, an ADT module is needed simultaneously by other modules, thus causing contention. To resolve this, we clone the module instance in demand and distribute the copies across different processors, so …

Wright State University College Of Engineering And Computer Science Bits And Pcs Newsletter, Volume 10, Number 5, May 1994, College Of Engineering And Computer Science, Wright State University

BITs and PCs Newsletter

A twelve page newsletter created by the Wright State University College of Engineering and Computer Science that addresses the current affairs of the college.

Load Balancing The Heat Equation In A Heterogeneous Environment With Pvm, N. Nemer-Preece, L. Mullin

Computer Science Technical Reports

Parallel processing is maturing with the ability to program heterogeneous environments and the help of networking systems such as PVM. Heterogeneous multiprocessing has been revolutionized due to tools such as PVM which allows the user to develop code independent of the machine arthitecture. This thesis develops a load balancing algorithm to handle the time march problem, heat conduction. It is based on the speeds of the machines in the current environment.

The heat equation algorithm is enhanced by the Psi Calculus of a Mathematics of Arrays. \l' Calculus allows the indexing of the heat matrix in terms of starts, stops …

Providing Assurance For Reponsive Computing Systems, G. Tsai, B. Mcmillin

Computer Science Technical Reports

A responsive computing system is a hybrid of real-time, distributed and fault-tolerant systems. In such a system, severe consequences will occur if the logical and physical specifications of the system are not met. In this dissertation, an approach to ensure satisfaction of specifications in the operational environment is presented as follows. First, we specify properties of the system using ITL formulas. Next, we collect, at runtime, events and maintain equivalent event histories to represent system execution. Finally, we apply a decision procedure to determine satisfaction of the formulas. A run-time evaluation system was built according to this approach and we …

Skin Cancer Diagnosis Using Hierarchal Neural Networks And Fuzzy Logic, H. C. Lee, F. Ercal

Computer Science Technical Reports

Skin cancer, the most common cancer in the United States that affects about 600,000 Americans every year, accounts for 1% of all cancer deaths. Among all, Malignant Melanoma is the most virulent form of skin cancer that is responsible for 75% of all deaths from skin cancer. In 1992, approximately 32,000 people are expected to develop melanoma and about 6,700 will die. However, even malignant melanoma can be treated successfully if detected in the early phase. Therefore, our research goal is to diagnose skin cancer, especially malignant melanoma.

In this study, only the digitized images obtained from color tumor slides …

Classification Characteristics Of Som And Art2, J. Aleshunas, D. St. Clair, W. Bond

Computer Science Technical Reports

Artificial neural network algorithms were originally designed to model human neural activities. They attempt to recreate the processes involved in such activities as learning, short term memory, and long term memory. Two widely used artificial neural network algorithms are the Self-Organizing Map (SOM) and the Adaptive Resonance Theory (ART2). Each was designed to simulate a particular biological neural activity. Both can be used as unsupervised data classifiers.

This paper compares performance characteristics of two unsupervised artificial neural network architectures; the SOM and the ART2 networks. The primary factors analyzed were classification accuracy, sensitivity to data noise, and sensitivity to the …

The Myriad Federated Database Prototype, San-Yih Hwang, Ee Peng Lim, H. R. Yang, S. Musukula, K. Mediratta, M. Ganesh, D. Clements, J. Stenoien, Jaideep Srivastava

Research Collection School Of Computing and Information Systems

Myriad is a federated database system (FDBS) prototype being developed at the University of Minnesota. The main objective behind this prototyping effort is to provide "enterprise-wide" information by integrating independently developed databases while preserving the local autonomy of the component DBMSs and applications. In Myriad, multiple federations can be formed. A federation consists of an integrated database whose schema is represented as a set of integrated relations derived from the export relations provided by the component DBMSS. SQL, mainly due to its simplicity and popularity among database users and vendors, has been adopted to express global queries as well as …

Managing Memory For Real-Time Queries, Hwee Hwa Pang, Michael J. Carey, Miron Livny

Research Collection School Of Computing and Information Systems

The demanding performance objectives that real-time database systems (RTDBS) face necessitate the use of priority resource scheduling. This paper introduces a Priority Memory Management (PMM) algorithm that is designed to schedule queries in RTDBS. PMM attempts to minimize the number of missed deadlines by adapting both its multiprogramming level and its memory allocation strategy to the characteristics of the offered workload. A series of simulation experiments confirms that PMM's admission control and memory allocation mechanisms are very effective for real-time query scheduling.

Subgrouped Real Time Recurrent Learning Neural Networks, Jeffrey S. Dean

Theses and Dissertations

A subgrouped Real Time Recurrent Learning (RTRL) network was evaluated. The one layer net successfully learns the XOR problem, and can be trained to perform time dependent functions. The net was tested as a predictor on the behavior of a signal, based on past behavior. While the net was not able to predict the signal's future behavior, it tracked the signal closely. The net was also tested as a classifier for time varying phenomena; for the differentiation of five classes of vehicle images based on features extracted from the visual information. The net achieved a 99.2% accuracy in recognizing the …

Knowledge-Based Nonuniform Crossover, Harpal Maini, Kishan Mehrotra, Chilukuri K. Mohan, Sanjay Ranka

Electrical Engineering and Computer Science - Technical Reports

We present a new "knowledge-based non-uniform crossover" (KNUX) operator for genetic algorithms (GA's) that generalizes uniform crossover. We extend this to "Dynamic KNUX" (DKNUX), which constantly updates the knowledge extracted so far from the environment's feedback on previously generated chromosomes. KNUX can improve on good solutions previously obtained by using other algorithms. The modifications made by KNUX are orthogonal to other changes in parameters of GA's, and can be pursued together with any other proposed improvements. Whereas most genetic search methods focus on improving the move-selection procedures, after having chosen a fixed move-generation mechanism, KNUX and DKNUX make the move-generation …

Classification Characteristics Of Som And Art2, J. J. Aleshunas, Daniel C. St. Clair, William E. Bond

Mathematics and Statistics Faculty Research & Creative Works

Artificial neural network algorithms were originally designed to model human neural activities. They attempt to recreate the processes involved in such activities as learning, short term memory, and long-term memory. Two widely used unsupervised artificial neural network algorithms are the Self-Organizing Map (SOM) and Adaptive Resonance Theory (ART2). Each was designed to simulate a particular biological neural activity. Both can be used as unsupervised data classifiers. This paper compares performance characteristics of two unsupervised artificial neural network architectures; the SOM and the ART2 networks. The primary factors analyzed were classification accuracy, sensitivity to data noise, and sensitivity of the algorithm …

Generating Indexing Functions Of Regularly Sparse Arrays For Array Compilers, Scott Thibault, Lenore Mullin, Matt Insall

Computer Science Technical Reports

There are many applications involving arrays that contain non-zero components in regular geometric partitions. These include triangular, diagonal, tridiagonal, banded, etc. When computing with this type of arrays, they are usually stored in a packed form and computations are performed with only the non-zero components. This packed form requires an indexing function that maps an index of the array to an index of the packed lexico-graphically stored array. This paper presents a method of describing regular partitions and of automatically generating an indexing function from that description. These methods enable an array compiler to compile array operations on these type …

Authorship Analysis: Identifying The Author Of A Program, Ivan Krsul

Department of Computer Science Technical Reports

No abstract provided.

A Multi-Chip Module Implementation Of A Neural Network, Tony R. Martinez, George L. Rudolph, Linton G. Salmon, Matthew G. Stout

Faculty Publications

The requirement for dense interconnect in artificial neural network systems has led researchers to seek high-density interconnect technologies. This paper reports an implementation using multi-chip modules (MCMs) as the interconnect medium. The specific system described is a self-organizing, parallel, and dynamic learning model which requires a dense interconnect technology for effective implementation; this requirement is fulfilled by exploiting MCM technology. The ideas presented in this paper regarding an MCM implementation of artificial neural networks are versatile and can be adapted to apply to other neural network and connectionist models.

Decoding Linear Block Codes Using A Priority-First Search: Performance Analysis And Suboptimal Version, Yunghsiang S. Han, Carlos R.P. Hartmann, Kishan Mehrotra

Electrical Engineering and Computer Science - Technical Reports

An efficient maximum-likelihood soft-decision decoding algorithm for linear block codes using a generalized Dijkstra's Algorithm was proposed by Han, Hartmann, and Chen. In this report we prove that this algorithm is efficient for most practical communication systems where the probability of error is less than 10-3 by finding an upper bound of the computation performance of the algorithm. A suboptimal decoding algorithm is also proposed. The performance of this suboptimal decoding algorithm is within 0.25 dB and 0.5 dB of the performance of an optimal decoding algorithm for the (104, 52) binary extended quadratic residue code and the (128, 64) …

Proof Of Correctness For Asocs Aa3 Networks, J. Cory Barker, Tony R. Martinez

Faculty Publications

This paper analyzes adaptive algorithm 3 (AA3) of adaptive self-organizing concurrent systems (ASOCS) and proves that AA3 correctly fulfills the rules presented. Several different models for ASOCS have been developed. AA3 uses a distributed mechanism for implementing rules so correctness is not obvious. An ASOCS is an adaptive network composed of many simple computing elements operating in parallel. An ASOCS operates in one of two modes: learning and processing. In learning mode, rules are presented to the ASOCS and incorporated in a self-organizing fashion. In processing mode, the ASOCS acts as a parallel hardware circuit that performs the function defined …