The overall student enrollment has declined over the past eight years in US. The administrators in enrollment management are applying ever more sophisticated analytical techniques in efforts to determine the characteristics of students that are most probable to enroll upon offering them monetary incentives (i.e., financial aid and/or scholarship). In this paper we present a framework that reveals the characteristics of such students using different machine learning models. Particularly, we use a genetic algorithm optimization method along with three different classification models: logistic regression (LR), support vector machines (SVMs) and bayesian networks (BNs). The results show that students with particular gender, ethnicity, socioeconomic and academic backgrounds are most probable to enroll upon offering them institutional money. We validated our results using data of actual students from a large public Tier 1 research university.

In almost every engineering curriculum, an introductory engineering course to enlighten students knowledge on the different engineering disciplines is offered for the freshman students. To better describe the engineering majors and increase students interest, practical projects are often requested to complement the theoretical lectures and concepts covered in these classes. In here we present a robotic-based design project that is designed and developed to serve as one of several final projects available for an introduction to engineering class. The projects require minimal-to-no previous knowledge in programming and attempts to give the students an enjoyable and rewarding experience while being introduced to the computer engineering discipline. The survey results clearly show that the projects enhance students' motivation and improve their interpretation to the engineering design process.

A novel theory was recently reported for the avalanche multiplication process in avalanche photodiodes (APDs) under dynamic reverse-biasing conditions. It has been shown theoretically that the bit-synchronized, periodic modulation of the electric field in the multiplication region can offer improvements in the gain-bandwidth product by reducing intersymbol interference in optical receivers. This paper reports a rigorous formulation of the sensitivity of optical receivers that employ dynamically biased APDs. To enable the sensitivity analysis, a recurrence theory is developed to calculate the joint probability distribution function (PDF) of the stochastic gain and avalanche buildup time in APDs that are operated under dynamic biasing. It is shown that in an ideal buildup-time limited scenario, a minimum receiver sensitivity of -20 dBm is predicted at an optimal gain of approximately 47 for a 60 Gb/s communication system, compared to a minimum of 0 dBm in the static-bias case. The receiver sensitivity analysis also reveals that, as the peak-to-peak voltage of the dynamic reverse bias increases, the device optimal gain increases while maintaining a short avalanche buildup time and reduced ISI. Of course, a point of diminishing return exists in practice when the tunneling current in the multiplication region becomes dominant.

We propose a new architecture for dynamic spectrum sharing called the distributed dynamic spectrum leasing (D-DSL) and a game-theoretic framework for its implementation on a cognitive radio network. In D-DSL, each available frequency channel is assigned to a primary user who may independently lease the channel to secondary users. Secondary users are allowed to transmit in multiple channels simultaneously. We establish conditions for the system to reach an equilibrium and analyze the robustness of the proposed game theoretic D-DSL in time-varying channels. For the same amount of available primary spectrum, D-DSL leads to better overall performance compared to previously proposed Centralized-DSL.

In this paper, we propose the concept of centralized dynamic spectrum leasing (C-DSL), in which multiple primary users belonging to the same primary system participate in the spectrum leasing process with secondary users (potential bidder for spectrum) under centralized control. We develop a new game-theoretic user interaction model suitable for C-DSL in a cognitive radio network. Dynamic Spectrum Leasing (DSL), proposed in allows active participation of both primary and secondary users in the spectrum sharing process. Motivated by network spectrum utilization considerations, we propose generalizations to the primary system utility function defined in and a new utility function for the secondary users. We also generalize the proposed non-cooperative C-DSL game to allow for linear multiuser detectors at the secondary base stations. We formulate the conditions on the primary system and the secondary user utility functions so that the proposed C-DSL game has desired equilibrium properties. We prove that the proposed C-DSL game has unique Nash equilibria (NE) under both matched filter (MF) and linear minimum mean-squared error (LMMSE) receivers. Equilibrium performance of the system and robustness of the proposed game theoretic adaptive implementation are investigated through simulations.