1,721,020 research outputs found
A Numerical Procedure for the 3D Design and the Optimisation of Active Shield Configurations for Stationary Magnetic Fields
An Efficient Hybrid Algorithm for the Optimisation of Problems with several Local Minima
Algorithm for the Fast Calculation of Magnetic Fields Generated by Arc-Shaped Conductors with Rectangular Cross Section
This paper presents a novel procedure for the fast numerical integration of magnetic field produced by arc-shaped conductors, characterized by rectangular cross section. Available procedures are based on the analytical integration of Biot-Savart’s law but most of them exploit the analytical integration along one coordinate, and then perform a two-dimensional numerical integration. In the proposed procedure, the analytical integration was performed along two coordinates, obtaining a one-dimensional integrand (thus avoiding the use of Elliptic Integrals), very easy to process using a state-of-the-art numerical quadrature library. The result is very satisfactory in terms of high speed and precision, particularly on conductor surface, and when its cross dimensions are very uneven
"Response Surface" method in magnetic optimization
The use of "response surface" method in the sensitivity analysis and optimisation of electromagnetic devices is presented. The algorithm proposed is based on the first and second order polynomial least square fitting of the objective function of the optimisation. The choice between the first and second order model is made by looking at the local behaviour of the function. When needed, a complete second order model is built and an analysis of the approximating function is performed to put in evidence the possible presence of "valleys" that can be followed to get the minimum. An outline of the strategy proposed and some results obtained both with analytically and numerically evaluated objective functions are presented and discussed
An Improved Simulated Annealing Algorithm for the Channel Allocation Problem in cellular networks
Appropriate channel allocation on each cell is one
of the most challenging issues in setting up and operating a
cellular system, both for its practical importance, as the correct
reuse of the available frequencies is crucial to guarantee good
quality of service and good operating conditions of filters and
power amplifiers, and its inherent difficulty, being the Channel
Allocation Problem (CAP) combinatorial, NP-hard and very
highly constrained. In this paper an improvement to a Simulated
Annealing based algorithm, in conjunction with several
heuristics, is presented to face the problem, and its efficiency in
solving a real-world application is shown
Fast Calculation of Magnetic Fields Produced by Rectangular Cross Section, Arc-Shaped Conductors
This paper presents a novel procedure for the fast numerical integration of magnetic field produced by arc-shaped conductors, characterized by rectangular cross section. Available procedures are based on the analytical integration of Biot-Savart’s law but, most of them, exploit the analytical integration along one coordinate, and then perform a two dimensional numerical integration. In the proposed procedure, the analytical integration was performed along two coordinates, obtaining a one dimensional integrand (thus avoiding the use of Elliptic Integrals), very easy to process using a state-of-the-art numerical quadrature library. The result is very satisfactory in terms of high speed and precision, particularly on conductor surface, and when its cross dimensions are very uneven
Three dimensional Finite Element Analysis in the Design of an AC-DC Converter for Arc Furnace feeding
A Simulated Annealing based Algorithm for the Optimisation of the Channel Allocation Problem in GSM networks
One of the most complex tasks in setting up a GSM
network is the appropriate choice of the set of channels to be used on
each BTS (Base Transceiver Station, or “site”), so as to minimize as far
as possible the interference inherently caused by the reuse of the same
channels in other cells. The Channel Allocation Problem (CAP) is
combinatorial, very highly constrained (particularly due the Italian RF
regulations), and NP-complete. An optimizer based on the Simulated
Annealing technique was developed, and used in conjunction with
several heuristics, in order to make it efficient and able to give good
frequency plans
- …
