Learning the global maximum with parameterized learning automata

A feedforward network composed of units of teams of parameterized learning automata is considered as a model of a reinforcement teaming system. The internal state vector of each learning automaton is updated using an algorithm consisting of a gradient following term and a random perturbation term. It is shown that the algorithm weakly converges to a solution of the Langevin equation implying that the algorithm globally maximizes an appropriate function. The algorithm is decentralized, and the units do not have any information exchange during updating. Simulation results on common payoff games and pattern recognition problems show that reasonable rates of convergence can be obtained.

Mechanism of seizure of aluminium-silicon alloys dry sliding against steel

A steel ball was slid on aluminium-silicon alloys at different temperatures. After the coefficient of friction had been measured, the surface shear stress was deconvoluted using a two-term model of friction. The ratio of surface shear stress to bulk hardness was calculated as a function of temperature, silicon content and alloying additions. These results are qualitatively similar to those recorded for pre-seizure specimens slid against an En24 disc in a pin-on-disc machine. This similarity, when viewed in the context of the phenomenon of bulk shear, provides a model for seizure of these alloys.

Suboptimal Midcourse Guidance of Interceptors for High-Speed Targets with Alignment Angle Constraint

Using the recently developed computationally efficient model predictive static programming and a closely related model predictive spread control concept, two nonlinear suboptimal midcourse guidance laws are presented in this paper for interceptors engaging against incoming high-speed ballistic missiles. The guidance laws are primarily based on nonlinear optimal control theory, and hence imbed effective trajectory optimization concepts into the guidance laws. Apart from being energy efficient by minimizing the control usage throughout the trajectory (minimum control usage leads to minimum turning, and hence leads to minimum induced drag), both of these laws enforce desired alignment constraints in both elevation and azimuth in a hard-constraint sense. This good alignment during midcourse is expected to enhance the effectiveness of the terminal guidance substantially. Both point mass as well as six-degree-of-freedom simulation results (with a realistic inner-loop autopilot based on dynamic inversion) are presented in this paper, which clearly shows the effectiveness of the proposed guidance laws. It has also been observed that, even with different perturbations of missile parameters, the performance of guidance is satisfactory. A comparison study, with the vector explicit guidance scheme proposed earlier in the literature, also shows that the newly proposed model-predictive-static-programming-based and model-predictive-spread-control-based guidance schemes lead to lesser lateral acceleration demand and lesser velocity loss during engagement.

Stimulation by DIF causes an increase of intracellular Ca2+ in Dictyostelium discoideum

Using fluorescence- activated cell sorting (FAGS), we have studied the effect of the differentiation-inducing factor (DIF) on cellular Ca2+ in Dictyostelium discoideum. We have shown previously that freshly starved or postaggregation amoebae are heterogenous with respect to the amounts of cellular Ca2+ that they contain; the L or ''low Ca2+'' class exhibits a prespore tendency and the H or ''high Ca2+'' class exhibits a prestalk tendency. Upon adding DIF, within 2 min there is an approximately twofold increase in the relative fraction of amoebae falling in the H class. A major part of the increase is caused by Ca2+ influx from the extracellular medium. Therefore a rise in the level of cellular Ca2+ is an early step in the signal transduction pathway following stimulation by DIF. Also, in parallel with the cellular heterogeneity in respect of Ca2+ content, there is a heterogeneity in the response to DIF, which appears to be restricted to L cells. (C) 1997 Academic Press.

Thermal sensitivity analysis of spacecraft battery

Transient thermal sensitivity is studied for systems that are subjected to conductive heat transfer within themselves and radiative heat transfer with the surrounding environment, including solar heat radiation, The battery in the Indian national communication satellite is one such system for which the studies are conducted with respect to panel conduction, conductance of insulating blanket, power dissipation within the battery, and absorptance and emittance of various elements, Comparison of sensitivities revealed that battery temperature is sensitive to its power dissipation during the beginning of life of the spacecraft, whereas toward the end of life of the spacecraft mission, the effect of absorptance of optical solar reflector is dominating, The influence of optical property values of the multilayer insulation blanket is almost negligible. Among the parameters studied in this analysis, the battery temperature is found to be mast sensitive to emittance of the optical solar reflector.

Synthesis of nonlinear controller to recover an unstable aircraft from poststall regime

Dynamics of the aircraft configuration considered in this paper show a unique characteristic in that there are no stable attractors in the entire high angle-of-attack flight envelope. As a result, once the aircraft has departed from the normal flight regime, no standard technique can be applied to recover the aircraft. In this paper, using feedback linearization technique, a nonlinear controller is designed at high angles of attack, which is engaged after the aircraft departs from normal flight regime. This controller stabilizes the aircraft into a stable spin. Then a set of synthetic pilot inputs is applied to cause an automatic transition from the spin equilibrium to low angles of attack where the second controller is connected. This controller is a normal gain-scheduled controller designed to have a large domain of attraction at low angles of attack. It traps the aircraft into a low angle-of-attack level flight. This entire concept of recovery has been verified using six-degrees-of-freedom nonlinear simulation. Feedback linearization technique used to design a controller ensures internal stability only if the nonlinear plant has stable zero dynamics. Because zero dynamics depend on the selection of outputs, a new method of choosing outputs is described to obtain a plant that has stable zero dynamics. Certain important aspects pertaining to the implementation of a feedback linearization-based controller are also discussed.

Health monitoring,of a helicopter rotor in forward flight using fuzzy logic

A fuzzy logic system is developed for helicopter rotor system fault isolation. Inputs to the fuzzy logic system are measurement deviations of blade bending and torsion response and vibration from a "good" undamaged helicopter rotor. The rotor system measurements used are flap and lag bending tip deflections, elastic twist deflection at the tip, and three forces and three moments at the rotor hub. The fuzzy logic system uses rules developed from an aeroelastic model of the helicopter rotor with implanted faults to isolate the fault while accounting for uncertainty in the measurements. The faults modeled include moisture absorption, loss of trim mass, damaged lag damper, damaged pitch control system, misadjusted pitch link, and damaged flap. Tests with simulated data show that the fuzzy system isolates rotor system faults with an accuracy of about 90-100%. Furthermore, the fuzzy system is robust and gives excellent results, even when some measurements are not available. A rule-based expert system based on similar rules from the aeroelastic model performs much more poorly than the fuzzy system in the presence of high levels of uncertainty.

Qualitative theorem proving in linear constraints

We know, from the classical work of Tarski on real closed fields, that elimination is, in principle, a fundamental engine for mechanized deduction. But, in practice, the high complexity of elimination algorithms has limited their use in the realization of mechanical theorem proving. We advocate qualitative theorem proving, where elimination is attractive since most processes of reasoning take place through the elimination of middle terms, and because the computational complexity of the proof is not an issue. Indeed what we need is the existence of the proof and not its mechanization. In this paper, we treat the linear case and illustrate the power of this paradigm by giving extremely simple proofs of two central theorems in the complexity and geometry of linear programming.