Deng Pan

We consider the challenge of resource allocation for a multi-user mobile edge computing system with multiple computing access points (CAPs). The user has to optimally allocate its tasks to either the CAP and cloud, or process them locally. This offloading game can be formulated as a non-convex centralized optimization problem and solved with a game theoretic approach where we prove that selfish mobile users is willing to participate in and truthfully follow the decisions of the FIP algorithm devised. We then evaluate the communication and computational costs at a Nash Equilibrium along with the mean random placement of offloading strategies for each user at various parameter settings.

The phenomenal increase in transistor scaling, known as Moore's Law, has powered advances in computing for over 50 years. Moore's Law has come to end with a dramatic slowing of scaling due to physical difficulties of semiconductors. The 3D FinFET is highly promising to enable continued scaling in emerging IC systems for intelligent sensory applications. We present a computational model to simulate SNM and WVM changes in presence of prohibitive device variations and detrimental short channel effects on SRAM cell memory circuits in 7nm FinFET technology. With the help of Monte Carlo methods, we introduce random threshold voltage fluctuations drawn from Gaussian distributions.

Dielectric elastomers function on the basis of a voltage-induced deformation for purposes of energy actuation or generation. We focus on the uniaxial prestrain dependence of VHB 4905 elastomer at room temperature. Data collected from a frequency analyzer showed that the permittivity decreases with increasing prestretch and frequency. We apply a modified version of the Cole-Cole numerical model to describe the dielectric constant of VHB 4905 elastomer accounting for an additional temperature variable. The results showed that the adoption of a product law with polynomials into the Cole equation surpasses the previous Debye model as the latter is only accurate in high-frequency ranges.

We examine the effects of an autocorrelation-based adaptive loop gain controller (ALGC) algorithm used for plesiochronous synchronization inside clock and data recovery circuits in the multi-gigahertz frequency domain. As data is transmitted over a lossy wireline channel, the signals propagating will inevitably be corrupted by a form of distortion known as jitter. It is the objective of the CDR to detect and mitigate such jitter at the receiver end of the channel during the process of recovering the transmitted data and its timing information. The proposed dual-parameter based ALGC algorithm can optimally minimize the output jitter variance in simulations while introducing key control advantages.