Differential Evolution for RFID Antenna Design
James Montgomery
ARTIFICIAL INTELLIGENCE SEMINARDATE: 2012-05-25
TIME: 15:00:00 - 16:00:00
LOCATION: RSISE Seminar Room, ground floor, building 115, cnr. North and Daley Roads, ANU
CONTACT: JavaScript must be enabled to display this email address.
ABSTRACT:
From the staff card in your wallet (or around your neck) to luggage passing through Hong Kong International Airport, radio-frequency identification (RFID) has become an essential technology for uniquely identifying items and people. An essential part of RFID is the convoluted, space-filling antenna belonging to the RFID tag embedded in a card or printed onto a sticker. The shape of the antenna affects its read distance, with larger distances (for smaller power inputs) preferred. Although these antennas have typically been designed by skilled engineers, recent attempts have been made to pose the design problem as a combinatorial optimisation problem. One of the first approaches used the constructive metaheuristic ant colony optimisation (ACO), as the problem of choosing an antenna path is naturally constructive. This seminar examines the application of the differential evolution algorithm to this problem and answers the question: how can a heuristic for optimisation in continuous spaces be sensibly adapted to solve the inherently discrete problem of laying out an antenna?
Questions and digressions will be welcome during the presentation. Possible
topics for further discussion include: how might DE be applied to my pet
problem; how and why would we apply a continuous solver to a discrete
problem; and multiobjective optimisation, including the many metrics for
evaluating algorithm performance.
