Simultaneous tuning of metaheuristic parameters for various computing budgets
Juergen Branke, Jawad Elomari · 2011
Many heuristics require a number of parameters to be tuned. One way to do this is meta-optimization: a higher level heuristic searches for the best parameter settings of a lower level heuristic which solves the optimization problem. However, the optimal parameter settings depend on the computational budget or running time available to the lower level heuristic. In this paper, we present a new meta-optimization approach to identify the best parameter settings simultaneously for various computational budgets.