SELECTING OBSERVATIONS FOR THE SPOT5 SATELLITES VIA TUNED GENETIC ALGORITHMS

This article considers the problem of selecting a subset of n photographs from a candidate set for maximizing a profit function. The problem considers a large number of imperative constraints reflecting the technological capabilities for the French earth observing satellites such as SPOT5. To solve the problem, a binary Genetic Algorithm (GA) way designed by using an outer level GA with a genome representation consisting of genes holding alleles which represent the inner level algorithm's parameters. These parameters were selection, crossover, mutation operators, crossover and mutation probabilities, and replacement percentage. The GA was able to identify solutions with higher objective value than the most recently published results in the SPOT5 literature.