Husam Hamad

Space partitioning in engineering design via metamodel acceptance score distribution

Metamodels for simulators are used to reduce computational costs in engineering system design. In general, metamodels have different fit accuracy levels over different regions in the design variables space, especially for highly nonlinear responses over wide spaces. A metamodeling strategy should place less emphasis on sub regions in the design variables space requiring relatively less complex metamodels. In this paper, we present a graphically based methodology that can be used to partition the space for piecewise metamodel building. The method is based on analyzing an initial global metamodel for acceptability in terms of prediction accuracy over the whole space; then generating acceptance score distribution (ASD) plots. Visual inspection of the ASD plots is used as a guidance to partition the design variables space, leading to a metamodel with fair prediction accuracy on a piece-by-piece basis, not just globally on average. The proposed methodology is general and can be applied to various metamodel types. It is tested on a number of problems, including some of the most highly nonlinear test problems used in the literature.

A new metric for measuring metamodels quality-of-fit for deterministic simulations

Metamodels are used to provide simpler prediction means than the complex simulation models they approximate. Accuracy of a metamodel is one fundamental criterion that is used as the basis for accepting or rejecting a metamodel. Average-based metrics such as root-mean-square error RMSE and R-square are often used. Like all other average-based statistics, these measures are sensitive to sample sizes unless the number of test points in these samples is adequate. A new metric that can be used to measure metamodels fit quality, called metamodel acceptability score MAS, is introduced. The proposed metric gives readily interpretable meaning to metamodels acceptability. Furthermore, initial studies show that MAS is less sensitive to test sample sizes compared to average-based validation measures

Discovering metamodels' quality-of-fit for simulation via graphical techniques

Abstract Metamodels are used in many disciplines to replace simulation models of complex multivariate systems. To discover metamodels ‘quality-of-fit’ for simulation, simple information returned by average-based statistics, such as root-mean-square error RMSE, are often used. The sample of points used in determining these averages is restricted in size, especially for simulation models of complex multivariate systems. Obviously, decisions made based on average values can be misleading when the sample size is not adequate, and contributions made by each individual data point in such samples need to be examined. This paper presents methods that can be used to discover metamodels quality-of-fit graphically by means of two-dimensional plots. Three plot types are presented; these are the so-called circle plots, marksman plots, and ordinal plots. Such plots can be used to facilitate visual inspection of the effect on metamodel accuracy of each individual point in the data sample used for metamodel validation. The proposed methods can be used to complement quantitative validation statistics; in particular, for situations where there is not enough validation data or the validation data is too expensive to generate.

Two new subjective validation methods using data displays

Three graphical data displays of histograms, box plots, and behavior plots are used in existing literature for subjective model validation. In this paper, we present two additional plots that can be used for displaying graphs of data; these are the so-called circle plots and ordinal plots. These plots are easy to generate using model data and system data. Like the existing plot types, no statistical assumptions are made on the data that are represented. However, more expeditious subjective interpretations about model operational validity are made using the methods presented

Subjective validation methods for analog integrated circuits’ metamodels using graphical displays of data

Metamodels are approximations to simulation models. They are built and validated using simulation results for samples of data points in the design variables space. Metamodels are more efficient to run compared to the simulation models they represent. They are validated to substantiate their accuracy using objective and/or subjective techniques. Objective validation methods based on various statistics such as root mean square RMS errors are often used. These methods require that certain statistical assumptions be satisfied by the data used in validation. Subjective validation methods are used, in particular, when some of these statistical assumptions are violated; for example, if the number of data points used is not sufficient. This paper presents and compares six different plot types that can be used to display data for subjective validation of metamodels, and demonstrates their usefulness as validation tools for analog circuits’ metamodels. These plots are easy to generate, using metamodel and simulation data. Furthermore, these methods are general and require no statistical assumptions for the data that can be displayed.

Telemetry Design of a Vital Sign Recording System

Blood pressure, respiratory rate, body temperature, and pulse rate are vital signs that under certain pathological conditions require continuous monitoring. In this paper we present a novel design of a system that embeds these signals into a single waveform that can be transmitted without the need for time or frequency division multiplexing. The system depends on changing the frequency of a square wave oscillator. One signal with low frequency contents controls the On-time of the oscillator while the Off-time is controlled by another signal. The third and the fourth signals are used to control voltage controlled oscillators. The voltage controlled oscillators outputs are used as fluctuations during the On-time and during the Off-time. The main advantage of this system is the reduction in circuit complexity in both transmitter and receiver with accurate recovery of the transmitted signals. The design of the proposed system is presented in this paper along with all the corresponding simulations.

Validation of metamodels in simulation: a new metric

Metamodels are used to provide more efficient predictions than the underlying simulation models do, but at the price of reduced prediction accuracy. Statistics used to quantify this prediction accuracy include the root-mean square error (RMSE), the coefficient of determination R-square, and the average absolute error (AAE). Such statistics depend on the average prediction accuracy over the validation sample; i.e., these metrics are sensitive to the size of the validation sample. This article, therefore, introduces a new metric, called the Model acceptability score (MAS). Preliminary results indicate that MAS is less sensitive to the validation sample size. The article focuses on deterministic simulation, which is used in various engineering disciplines, e.g., electronic engineering.

High-speed 980nm VCSELs with integrated distributed losses for mode control

Single fundamental mode, oxide-confined, polyimide planarized 980nm vertical cavity surface emitting lasers (VCSELs) with a multi-oxide layer (MOL) structure to increase oxide aperture diameter and maintain single-mode operation are fabricated and characterized. VCSELs with an 8μm active diameter and 16 mode suppression layers maintained single transverse mode operation under continuous wave (CW) condition with a side-mode suppression ratio (SMSR) of more than 32 dB at current densities up to 20 times threshold, which is five times higher than the previously reported value for similar devices with only 3 mode suppression layers, and exhibited a maximum 3-dB modulation frequency bandwidth of 13GHz. The threshold current and voltage were as low as 260μA and 1.45V, respectively, with a maximum optical power and slope efficiency of 1mW and 0.31W/A, respectively.

Minimum bias metamodels in engineering system design

Experimental designs are techniques used to determine combinations of design variables to generate models of engineering systems. When experiments are conducted using simulators to determine system responses, the resulting approximation to the simulator is called a metamodel or a surrogate model. Accuracy of metamodels is tightly related to experimental designs. Designs that minimise errors caused by metamodel fitting inadequacy - called bias errors - are known as minimum bias designs (MBDs). This paper presents techniques for generating MBDs for use in response surface metamodelling of engineering systems to obtain minimum bias metamodels. The resulting MBDs are compared to the more recent space-filling designs such as the Latin hypercube (LHC) samples. The paper also includes tables of second- to fourth-order MBDs for two- to five-dimensional spaces.