Improved Performance of Global Optimisation Methods for Inversion Problems in Underwater Acoustics

To invert for parameters of the seabed and the water column, global optimisation methods are used. These methods search for the best parameter combination, so as to make the difference between modelled and measured pressure fields as small as possible. Simulated annealing and genetic algorithms (GA) are commonly used as global optimisation algorithms for this inversion. Their efficiency depends on the tuning of the algorithm. For a bottom model with in the order of ten geo-acoustic parameters, a relatively new global optimisation method - Differential Evolution (DE) - is applied. We claim a ten times higher efficiency of DE, as it finds the optimal parameter combination with only one tenth of the number of model evaluations needed by GA. Apart from the tuning of the algorithms, the performance of DE in comparison with GA is investigated in terms of efficiency and robustness