A combinatorial genetic algorithm for the configuration of the 2dF/AAOmega spectrograph at the anglo-Australian observatory

To help unravel the structure of the universe, astronomers have developed systems which observe large clusters of objects at the same time. One such system is the 2-degree field spectrograph at the Anglo Australian Observatory. This system uses 400 fibres which can observe and measure the spectra of up to 400 astronomical objects in parallel. These optical fibres are positioned by a robotic arm. During placement, complex fibre entanglements often occur, reducing the overall efficiency of the system. This work aims to develop a combinatorial genetic algorithm which can help increase the efficiency of these systems by evolving sequences of fibre re-positionings. We present an integer-based representation developed for this GA which has the ability to automatically satisfy hard constraints during sequence seeding, mutation and recombination. We also present some initial results in re-configuring fibre placements. Categories and Subject Descriptors I.2.8 [Artificial Intelligence]: Problem...

• Type: Conference
• Year: 2006
• Where: GECCO
• Authors: Steven Manos, Geraint Lewis
• Keywords: 2-degree Field Spectrograph, Fibre, GECCO 2006, Genetic Algorithm, Optimization