The realization of the missing fourth element by Hewlett-Packard in 2008, the memristor, adds new promising technology that enables the continuing improvement of performance, power and cost of integrated circuits and keeping Moore's law alive. Memristor-based technology provides much better scalability, higher utilization when used as memory, and overall lower power consumption. This paper presents a detailed study of existing memristor modeling using Matlab simulations. We studied three different models to predict the behavior of the memristor device. We developed the Matlab algorithms for all models and analyzed them for their compatibility with the experimentally established characteristics of HP memristor, as well as their viability for use in memory circuits. We discussed all the difficulties with these models and adopt a modified model that gives more realistic description of a memristor device.
The significant gap between quantitative and qualitative understanding of cytoskeletal function is a pressing problem; microscopy and labeling techniques have improved qualitative investigations of localized cytoskeleton behavior, whereas quantitative analyses of whole cell cytoskeleton networks remain challenging. Here we present a method that accurately quantifies cytoskeleton dynamics. Our approach digitally subdivides cytoskeleton images using interrogation windows, within which box-counting is used to infer a fractal dimension (D-f) to characterize spatial arrangement, and gray value intensity (GVI) to determine actin density. A partitioning algorithm further obtains cytoskeleton characteristics from the perinuclear, cytosolic, and periphery cellular regions. We validated our measurement approach on Cytochalasin-treated cells using transgenically modified dermal fibroblast cells expressing fluorescent actin cytoskeletons. This method differentiates between normal and chemically disrupted actin networks, and quantifies rates of cytoskeletal degradation. Furthermore, GVI distributions were found to be inversely proportional to D-f, having several biophysical implications for cytoskeleton formation/degradation. We additionally demonstrated detection sensitivity of differences in D-f and GVI for cells seeded on substrates with varying degrees of stiffness, and coated with different attachment proteins. This general approach can be further implemented to gain insights on dynamic growth, disruption, and structure of the cytoskeleton (and other complex biological morphology) due to biological, chemical, or physical stimuli. (C) 2016 Wiley Periodicals, Inc.
A challenge in transmission-distribution coordination is how to quickly and reliably coordinate Distributed Energy Resources (DERs) across large multi-stakeholder Distribution Networks (DNs) to support the Transmission Network (TN), while ensuring operational constraints continue to be met within the DN. Here we propose a hierarchical feedback-based control architecture for coordination of DERs in DNs, enabling the DN to quickly respond to power set-point requests from the Transmission System Operator (TSO) while maintaining local DN constraints. Our scheme allows for multiple independently-managed areas within the DN to optimize their local resources while coordinating to support the TN, and while maintaining data privacy; the only required inter-area communication is between physically adjacent areas within the DN control hierarchy. We conduct a rigorous stability analysis, establishing intuitive conditions for closed-loop stability, and provide detailed tuning recommendations. The proposal is validated via case studies on multiple feeders, including IEEE-123 and IEEE-8500, using a custom MATLAB-based application which integrates with OpenDSS. The simulation results show that the proposed structure is highly scalable and can quickly coordinate DERs in response to TSO commands, while responding to local disturbances within the DN and maintaining DN operational limits.
The realization of the missing forth element by HP in 2008, the memristor, which was theorized by Leon Chau in 1971, adds new promising technology that enables the continuing improvement of performance, power and cost of integrated circuits and keeping Moore's law alive. Memristor-based technology provides much better scalability, higher utilization when used as memory, and overall lower power consumption. The goal of this paper is to study the read and write behavior of memristors using mathematical and SPICE simulations. This modeling will help us understand the behavior and the stability of the memristor state especially under the repeated read process. We show in this study that reading a memristor by a decaying voltage can balance the tradeoff between stable reading and fast writing.