Complex metaheuristics

Abstract

a b s t r a c tComplexity is a prevalent feature of numerous natural and artificial systems and as such has attracted much scientific interest in the last decades. The pursuit of computational tools capable of analyzing, modeling or designing systems exhibiting this complex nature -in which the properties of the system are not evident at the bottom level but emerge from its global structure -is a major issue. Metaheuristics can play here an important role due to its intrinsic adaptability and powerful optimization capabilities. In many regards, metaheuristics are also examples of complex systems since their behavior emanates from the orchestrated interplay of simpler algorithmic components. This bidirectional connection between metaheuristics and complex systems offers numerous avenues for fruitful research.

Misfit Landforms Imposed by Ill-Conditioned Inverse Parametric Problems

“…The goal of next sections is to establish a taxonomy of solution sets for (7) that corresponds to the local taxonomy of the IPP's ill-conditioning. Such a taxonomy will be convenient for studying the features of complex solving strategies (e.g., memetic strategies [7]) that combine a global stochastic search with various algorithms that improve the search results locally. The latter can be both stochastic or deterministic and can either speed up finding isolated local minimizers or approximate the shapes of areas on which the objective attains a locally minimal constant value.…”