Hybrid learning models to get the interpretability–accuracy trade-off in fuzzy modeling

One of the problems associated to linguistic fuzzy modeling is its lack of accuracy when modeling some complex systems. To overcome this problem, many different possibilities of improving the accuracy of linguistic fuzzy modeling have been considered in the specialized literature. We will call these approaches as basic refinement approaches. In this work, we present a short study of how these basic approaches can be combined to obtain new hybrid approaches presenting a better trade-off between interpretability and accuracy. As an example of application of these kinds of systems, we analyze seven hybrid approaches to develop accurate and still interpretable fuzzy rule-based systems, which will be tested considering two real-world problems.This work has been supported by the spanish cicyt project tic2002-04036-c05-01 (keel).     

Assessing rabbit and chicken tissue susceptibility to oxidation through the ferrous oxidation‐xylenol orange method

We used an induced version of the FOX method to assess the oxidative stability of meats. Formerly, this induced version only measured the lipid hydroperoxide (LHP) value once the ferrous oxidation‐xylenol orange (FOX) reaction had reached a plateau. However, samples could have similar final LHP values that were reached in different ways. Thus, there may be variations in the samples' susceptibility to oxidation that are not detected by a final measurement of the LHP value. If the LHP value is measured at different points in the FOX reaction, it might be possible to calculate parameters such as induction time, oxidation rate, maximum LHP value, time of maximum LHP value and area under the curve. Such parameters might provide deeper insight into the evolution of oxidation in each sample than a final measurement of the LHP value. However, the accuracy of the measurement of these parameters could depend on the number of LHP measurements that can be taken during the FOX reaction. This improvement in the FOX method for assessing sample susceptibility to oxidation was applied to meats and other tissues such as liver.     

Mineralization of flagella for nanotube formation

Flagella isolated from Salmonella typhimurium were used as a template for the formation of TiO₂ nanotubes, which was accomplished through recent advances in low temperature ceramic film processing. TiO₂ is precipitated onto or attracted to the flagellar surface in aqueous solution. The mechanism of mineral film formation is similar to that of biominerals in nacre, sea urchin spine, and sponge spicule. This process is advantageous over other mineralization processes, as no genetic modification of the flagella is necessary and the mineralization is done at low temperature in aqueous solution.     

Effect of synthesis parameters on the electrical conductivity of polypyrrole‐coated poly(ethylene terephthalate) fabrics

The effects of pyrrole, anthraquinone‐2‐sulphonic acid (AQSA) and iron(III) chloride (FeCl₃) concentrations, reaction time and temperature on the electrical conductivity of polypyrrole (PPy)—coated poly(ethylene terephthalate) (PET) fabrics were investigated. With an increase in both the AQSA and FeCl₃ concentrations, resistivity decreased to a point beyond which higher concentrations led to increased surface resistivity. Erosion of the polymer coating, in dynamic synthesis from continual abrasion, manifested as an exponential increase in the resistance of the coated textile substrate. This was not encountered in static synthesis conditions. Temperature affected the degree of surface and bulk polymerisation. The effect of polymerisation temperature on conductivity was negligible. Conductive polymer coating on textiles through chemical polymerisation enabled a smooth coherent film to encase individual fibres, which did not affect the tactile properties of the host substrate. The optimum FeCl₃/pyrrole and AQSA FeCl₃/pyrrole molar ratios were found to be 2.22 and 0.40 respectively. Copyright © 2003 Society of Chemical Industry     

An Interconnect Strategy for a Heterogeneous, Reconfigurable SoC

Editor's note:The diversification of the SoC market has increased the need for domain-specific programmable platforms that can be tailored to postfabrication products. This article presents a design case study on the interconnect for such a programmable platform, as part of the Morpheus project.—Li-Shiuan Peh, Princeton University     

Squeeze flow as a tool for developing optimized gypsum plasters

This paper presents a rheological investigation of pure gypsum (PG) and a commercial gypsum–lime–filler plaster (CP) using the modified Vicat apparatus and squeeze flow method. The samples were tested at several different intervals after manual or mechanical mixing. The results confirmed squeeze flow to be more sensitive in determining fresh paste behavior than the modified Vicat apparatus. PG set faster when prepared in mechanical mixer than when manually mixed. Conversely, the CP composition presented longer setting when mixed mechanically. The study also included the analysis of two ready-to-use polymer-based products for leveling and rendering (drywall joint compound – DJC; acrylic putty – AP) measured by squeeze flow and compared to the commercial composition.     

Adaptive memory programing for the robust capacitated international sourcing problem

The International Sourcing Problem consists of selecting a subset from an available set of potential suppliers internationally located. The selected suppliers must meet the demand for items from a set of plants, which are also located worldwide. Since the costs are affected by macroeconomic conditions in the countries where the supplier and the plant are located, the formulation considers the uncertainty associated with changes in these conditions. We formulate the robust capacitated international sourcing problem by means of a scenario-optimization approach. When dealing with uncertainty, one of the most common approaches in the literature is to formulate the problem via a set of possible scenarios, this is the approach we follow here. In this paper we propose a constructive method based on memory structures to solve this problem. The method is coupled with a local search procedure followed by a path relinking for improved outcomes. We propose an innovative mechanism to achieve a good balance between intensification and diversification in the search process. Moreover, our path-relinking implementation uses constructive neighborhoods for extrapolated relinking. The computational experimentation favors this method when compared with a recent tabu search approach.