1 This is my implementation of the "Multi-Level Single-Linkage" (MLSL)
2 algorithm for global optimization by random local optimizations (a
3 multistart algorithm with "clustering" to avoid repeated detection of
4 the same local minimum), modified to optionally use a Sobol'
5 low-discrepancy sequence (LDS) instead of pseudorandom numbers. See:
7 A. H. G. Rinnooy Kan and G. T. Timmer, "Stochastic global optimization
8 methods," Mathematical Programming, vol. 39, p. 27-78 (1987).
9 [ actually 2 papers -- part I: clustering methods (p. 27), then
10 part II: multilevel methods (p. 57) ]
14 Sergei Kucherenko and Yury Sytsko, "Application of deterministic
15 low-discrepancy sequences in global optimization," Computational
16 Optimization and Applications, vol. 30, p. 297-318 (2005).
18 It is under the same MIT license as the rest of my code in NLopt (see