Master of Science, University of Toledo, 2021, Engineering (Computer Science)

This research examines Echo State Network, a reservoir computer, performance using four different benchmark problems, then proposes heuristics or rules of thumb for configuring the architecture, as well as the selection of parameters and their values, which are applicable to problems within the same domain, to help serve to fill the ‘experience gap' needed by those entering this field of study. The influence of various parameter selections and their value adjustments, as well as architectural changes made to an Echo State Network, a powerful recurrent neural network configured as a reservoir computer, can be difficult to understand without experience in the field, and even some hyperparameter optimization algorithms may have difficulty adjusting parameter values without proper manual selections made first; therefore, it is imperative to understand the effects of parameters and their value selection on echo state network architecture performance for a successful build. Thus, to address the requirement for an extensive background in Echo State Network architecture, as well as examine how Echo State Network performance is affected with respect to variations in architecture, design, and parameter selection and values, a series of benchmark tasks representing different problem domains, including time series prediction, pattern generation, chaotic system prediction, and time series classification, were modeled and experimented on to show the impact on the performance of Echo State Network.

Committee: Gursel Serpen (Advisor); Kevin Xu (Committee Member); Joshua Stuckner (Committee Member); Lawrence Thomas (Committee Member) Subjects: Computer Engineering; Computer Science

Master of Science (MS), Ohio University, 1984, Electrical Engineering & Computer Science (Engineering and Technology)

Identification of synchronous machine stability parameters using a quasilinearization-least-square-error algorithm

Committee: Nasser Jaleeli (Advisor) Subjects:

Doctor of Philosophy, Case Western Reserve University, 1994, Physics

Extensive experimental studies of fluid hydrocarbons in the lubricating range of molar mass have been undertaken sometime ago by American Petroleum Institute Project 42, located in the Departments of Chemistry and Physics at Pennsylvania State University. In these studies systematic structural changes were introduced, so that the equation of state (e.o.s.) as well as the viscosities of linear paraffins, branched hydrocarbons, and various rings attached to n-alkanes tails are known. Hence this material became the basis for various semi-empirical or empirical structural correlations. We proceed here with the hole theory of Simha-Somcynsky (SS) which has proven quantitatively successful for low as well as high molar mass system and examine e.o.s. data. We demonstrate the success of the theory and obtain the characteristic volume (ν*), energy (varepsilon*) and flexibility (c) parameters as functions of chain length for the different structures. For the short chains in question these represent averages over the terminal and internal units. By suitable generalization of the SS theory developed for physical mixtures we decompose these averages into the individual group contributions. The accuracy of the numerical procedures employed is tested by back computations. Sometime ago A. Bondi developed structural rules for the computation of Van der Waals excluded group volumes. Interesting correlations between these and the above ν* values, defined for a 6-12 potential, are obtained. In the same way we examine correlations between D. W. Van Krevelen's and P. J. Hoftyzer's cohesive group energies and varepsilon* values.

Committee: Robert Simha (Advisor) Subjects: