Team tao

Overall Objectives
Scientific Foundations
Application Domains
New Results
Contracts and Grants with Industry
Other Grants and Activities

Section: New Results

Continuous Optimization

Participants : Anne Auger, Dimo Brockhoff, Zyed Bouzarkouna, Nikolaus Hansen, Ilya Loshchilov, Mohamed Jebalia, Raymond Ros, Marc Schoenauer, Olivier Teytaud, Fabien Teytaud.

Distributed optimization:

Within the OMD2 project we have studied the convergence rate scaling of ESs in the parallel setting as well as the optimal choice of parents and weights for recombination [78] . Application of CMA-ES parallelized for calibration of traffic simulation has been carried out [119] . A simple and effective modification of the self-adaptive evolutionary algorithm for the highly parallel case has been proposed [96] : the selection pressure is modified in a principled way to tackle the highly parallel case. Theoretical work has been achieved on the ultimate limits of the parallelization of Evolutionary Algorithms, including simple and efficient modifications of existing algorithms [97] .

New algorithms:

With the motivation of designing new robust local search algorithms, we have proposed to use derandomization by mirroring combined with a smart selection mechanism called sequential selection [52] . Implementation of the different mechanisms have been made within CMA-ES and benchmarked intensively showing the improvement brought by both mechanisms [31] , [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] , [29] , [30] . A thorough theoretical studies where we prove lower bounds and quantify convergence rates has been carried out [40] , [113] . Active CMA has been revisited and combined with the (1+1)-ES [25] and with weighted recombination [76] , [75] , [74] .


The benchmarking platform COCO has been further developed and extensively used (see e.g. previous point). Restart variants of CMA-ES [92] , [93] , [91] and NEWUOA [91] , [94] , [95] have been benchmarked. Results from 31 algorithms in 2009 have been compared [73] .

Optimization with meta-models and surrogate:

A modification enhancing CMA-ES with local meta-models has been proposed in [48] and applied to the well-placement problem [49] . A new algorithm coupling CMA-ES with a surrogate computed with support vector machine has been proposed [81] . We proposed a mixture cross-entropy optimizer and used it for merit function optimization in a Gaussian-process surrogate optimization framework [41] .

Multi-objective optimization:

Theoretical foundations of hypervolume based search algorithms for bi-objective problems have been published in [8] . Extension to three objective problems has been carried out [28] . Surrogate for multi-objective algorithms have been proposed [82] , [105] , [80] . A simple but effective improvement for step-size adaptation in MO-CMA-ES has been found [99] .


Logo Inria