Heuristic Optimisation of 2000+ dimensional tests

Identification of multidimensional tasks? solutions where parameters are thousands and more, seems to be an embarrassing challenge for modern computational systems in terms of hardware capacity and software abilities. Even more, currently is difficult to find convincing evidence for methods capable of resolving heterogeneous tasks without retuning of algorithms parameters. This presentation focuses on evaluation of 2000 dimensional heterogeneous real-value numerical tests. The aim of presented study is to extend the knowledge on how high dimensionality reflects on search space complexity, in particular to identify minimal time and minimal number of objective function evaluations required for reaching acceptable solutions with non-zero probability on tasks with 2000 dimensions. It aims also to assess computational limitations by comparison of 2000 and 2016 dimensional tasks evaluation where search process is limited to the same number of functions? evaluations. Achieved experimental results are summarised and analysed and could be used for further research and comparison.