# Computers.

Some Results on Synchronization in Colored Digraphs. Ranjan Chaudhuri, Eastern Michigan University

Let G=(V,E) be a strongly connected digraph in which the out-degree of each vertex is equal to two. A red-blue coloring of G is a coloring of its edges with the colors red(R) and blue(B) such that each vertex has an outgoing edge colored red(R) and the other outgoing edge colored blue(B). We define a string I belonging to {R,B}* to he an. instruction. For a vertex v in V, let l(v) denote the vertex w that is reachable from v if one follows the directed path labeled by the colors in I. We define two vertices vl and v2 to be synchronizable with respect to an instruction if I(vl)=l(v2)=w for some w in V. Some interesting results on synchronizing instructions and synchronizable states in colored digraphs are presented in this paper.

Reinforcement Learning: An Examination of Path Length Asymmetries in Markov Decision Processes. Nicholas Alexander Perry, University of Michigan--Flint

The current paper deals with Computer Science as it relates to robotics and Artificial Intelligence. It examines the Markov Decision Process in solving Spatial Navigation problems in an abstract maze. The Markov Decision Process has been an important tool for researchers in the held of Reinforcement Learning for years. One of the more significant downsides of using a Markov Decision Process is that certain assumptions that appear trivial are actually not valid assumptions under the Markov Model. The reasons behind this property arc not well understood. My research examines one such assumption: path reversibility. This paper will explore the path length asymmetries that result from applying conventional reinforcement learning methods based on the Markov Decision Process to a spatial navigation problem.

JavaScript Unleashed. Augustine Ikeji, Eastern Michigan University

JavaScript was introduced in the early 1990s as a language that provided user interactivity with web pages. Back then, it only tan on Netscape browsers. Its popularity has since widened and all major browsers such as Internet Explorer and Fire fox now support it. It is currently the most popular scripting language on the internet and its applications and features have also widened. JavaScript's primary application may be form validation, but it docs provide a wide range of applications and features such as recursion, exception handling and Objects similar to those found in mainstream languages like Java and C+ +.

This presentation will give an overview of JavaScript, and explore its advanced applications in areas such as support for dynamic web programming, timers, event handling, recursion, and the HTML Document Object Model.

Clustering using Genetic Programming. Samir Tout, Keane, Inc., Eastern Michigan University, and William Sverdlik, Eastern Michigan University

As databases continue to grow in size, efficient and effective clustering algorithms play a vital role in data mining applications. Practical clustering faces several challenges including: identifying clusters of arbitrary shapes, sensitivity to the order of input, dynamic determination of the number of clusters, outlier handling, processing speed of massive data sets, handling higher dimensions, and dependence on user-supplied parameters. Many studies have addressed one or more of these challenges. This research explores using a hybrid combination of genetic programming, multi-objective fitness with density, and data parallelism to address some of the above challenges. The genetic programming represents arbitrary cluster shapes and circumvents falling in local optima. The multi-objective fitness function helps locate and avoid outliers. furthermore, data parallelism is also explored whereby a master processor breaks down the dataset into subsets that are handled simultaneously by multiple slave processors. Slave processors report their results back to the master processor where they are consolidated into a final solution. Other aspects of the genetic programming are also explored, such as variations of the crossover, mutation, structural operators, and a newly introduced repair operator, that helps align the cluster detections along the data distribution. Some promising experimental results will be presented.

Online Statistical Analysis Resource Kit (Poster). Joshua Urbain, Morteza Marzjarani, Saginaw Valley State University

An interactive web page is developed to gather information from the user for analysis. The analysis would involve the collection of data, modify the data (where needed), then incorporate a statistical software such as SAS, SPSS, or Minitab into the Well page to perform the analysis of data.

This software was developed in the PHP platform. The reasons for selecting this scripting language are among others, mastering heavy load, ease of use, and portability.