A novel approach, which is based on artificial neural network (ANN) by backpropagation, for fast and trusted herbal ingredient discoveries, is proposed. It is fast, because different ANN modules can be executed in parallel, and the ANN results are trustworthy, because they can be verified by TCM domain experts in real clinical environments. The ANN is able to learn the relationship between herbal ingredients and the set of information given (e.g. symptoms and illnesses). The ANN output is called the relevance index (RI), which conceptually associates two TCM entities (e.g. U and V) in a 2-D or 3-D manner (D for dimension). RI is the quantified P(U∩V) part of P(U ∪ V) = P(U) + P(V) - P(U ∩ V), an IT (information technology) formalism in which P stands for probability. The interpretation of P(U ∩ V) adheres to TCM formalism(s).

In this paper we propose to apply dynamic buffer tuning to make pervasive medical consultation successful. In pervasive consultation the physician can answer patient requests anytime, anywhere. The physician carries a SFF (small form factor) mobile device such as a mobile phone, which provides the interface to interact with the rest of the distributed Pervasive Digital HealthCare System (PDHD) on the mobile Internet. The PDHD in this paper caters to Traditional Chinese Medicine (TCM) and therefore called TCM-PDHD or simply T-PDHD. The T- PDHD usability relies on various factors such as: a) correct information transfer, and b) fast system response. This paper focuses on the second factor and explores how the dynamic buffer tuning technique can be exploited to quicken system response.

The M3RT-based real-time traffic pattern detector proposed identifies the Internet traffic pattern on the fly. Firstly it determines if a time series aggregate is stationary. Secondly it confirms if the aggregate exhibits short-range dependence (SRD) or long-range dependence (LRD). Thirdly it detects if the smooth system operation has suddenly become irregular and chaotic. This detection is achieved by computing the instantaneous value of the Holder exponent that has a (0,1) range to accommodate different degrees of fractality. When the Holder exponent has wandered outside the (0,1) region, fractal breakdown has occurred. The capability of detecting such breakdowns by a real-time application enables it to avoid sudden failure. The Intel's VTune Performance Analyzer indicates the proposed model can be deployed in real time effectively. This feature is of importance to the reliability improvement of Web applications which run on the Internet

Abstract In this paper two novel intelligent buffer overflow controllers: thefuzzy logic controller(FLC) and thegenetic algorithm controller(GAC) are proposed. In the FLC the extant algorithmic PID controller (PIDC) model, which combines theproportional(P),derivative(D) andintegral(I) control elements, is augmented with fuzzy logic for higher control precision. The fuzzy logic divides the PIDC control domain into finer control regions. Every region is then defined either by a fuzzy rule or a ‘don't care’ state. The GAC combines the PIDC model with the genetic algorithm, which manipulates the parametric values of the PIDC as genes in a chromosome. The FLC and GAC operations are based on the objective function$\{0,\Delta\}^2$ . The principle is that the controller should adaptively maintain the safety margin$\Delta$ around the chosen reference point (represent by the ‘0’ of$\{0,\Delta\}^2$ ) at runtime. The preliminary experimental results for the FLC and GAC prototypes indicate that they are both more effective and precise than the PIDC. After repeated timing analyses with theIntel's VTune Performer Analyzer, it was confirmed that the FLC can better support real‐time computing than the GAC because of its shorter execution time and faster convergence without any buffer overflow. Copyright © 2005 John Wiley & Sons, Ltd.

Two novel expert dynamic buffer tuners/controllers, namely, the neural network controller (NNC) and the fuzzy logic controller (FLC) are proposed in this paper. They use soft computing techniques to eliminate buffer overflow at the user/server level. As a result they help shorten the end-to-end service roundtrip time (RTT) of the logical Internet transmission control protocol (TCP) channels. The tuners achieve their goal by maintaining the given safety margin Delta around the reference point of the {0,Delta}2objective function. Overflow prevention at the Internet system level, which includes the logical channels and their underlying activities, cannot shorten the service RTT alone. In reality, unpredictable incoming request rates and/or traffic patterns could still cause user-level overflow. The client/server interaction over a logical channel is usually an asymmetric rendezvous, with one server serving many clients. A sudden influx of simultaneous requests from these clients easily inundates the server's buffer, causing overflow. If this occurs only after the system has employed expensive throttling and overflow management resources, the delayed overflow rectification could lead to serious consequences. Therefore, it makes sense to deploy an independent user-level overflow control mechanism to complement the preventative effort by the system. Together they form a unified solution to effectively stifle channel buffer overflow

This paper has two original and important proposals: a) the self-similarity (S/sup 2/) traffic filter for real-time applications and b) the R/sup 2/-FLC (reconfigurable real-time fuzzy logic controller) for on-line dynamic buffer size tuning. Inclusion of the S/sup 2/ filter into the extant RTPD (real-time traffic detector) model creates the novel enhanced RTPD (ERTPD) framework. The R/sup 2/-FLC uses the ERPTD to detect the current traffic pattern of a logical channel on the fly and auto-tunes accordingly. The auto-tuning process, which is system reconfiguration in the R/sup 2/-FLC context, neutralizes the ill effects by changing traffic patterns on the accuracy and stability of the dynamic buffer size tuning process.

The aim of the novel adaptive fuzzy logic controller (FLC) proposed in this paper is to eliminate buffer overflow at the reception side of an Internet channel. The elimination improves channel reliability and leads to shorter service roundtrip time (RTT) because of reduced retransmission. The FLC combines the extant, less adaptive PID controller with fuzzy logic, which divides the original PID control domain into finer fuzzy regions. Then, every region is governed by a predefined fuzzy rule or simply a don't care state that requires no computation or parametric changes. The goal of the don't care states is to offset the FLC complexity so that the controller execution time is reduced. The FLC adaptively maintains the given safety margin /spl Delta/ of the chosen objective function {0, /spl Delta/}/sup 2/. The value for /spl Delta/ is normally a preset difference between the buffer length and the queue length. The repeated timing analyses for the FLC with the Intel's VTune performer analyzer confirms its capability for supporting real-time computing efficaciously.

In an ORC (object oriented real time computing) environment, it is important to reduce message retransmissions during message passing because these retransmissions cause significant time delay, and this makes it difficult to achieve the necessary timeliness. A cause of retransmission is message loss due to buffer overflow at the reception side. Here a P+I+D (P for proportional, I for integral and D for derivative) adaptive buffer control algorithm is proposed to prevent such possible overflow. The I control depends on the Convergence Algorithm (CA), which is a stable and efficient IEPM (Internet End-to-End Performance Measurement) tool that predicts the mean message roundtrip time (RTT) of a communication channel quickly and accurately. The P+I+D algorithm was tested under different conditions in a mobile ORC (MORC) environment where mobile agents collaborate freely over the Internet. The different test results confirm that the proposed P+I+D approach is indeed effective for preventing buffer overflow.