Invited Session Tue.2.H 3503

Tuesday, 13:15 - 14:45 h, Room: H 3503

Cluster 6: Derivative-free & simulation-based optimization [...]

New techniques for optimization without derivatives


Chair: Stefan Wild and Luís Nunes Vicente



Tuesday, 13:15 - 13:40 h, Room: H 3503, Talk 1

Margaret Wright
Defining non-monotone derivative-free methods


Non-monotone strategies in optimization avoid imposing a monotonicity requirement at every iteration with the goal of achieving rapid convergence from an alternative strategy over a longer sequence of iterations. We consider how to define non-monotone derivative-free methods in, broadly, this same spirit, especially in light of recent worst-case complexity results that are closely tied to monotonicity requirements.



Tuesday, 13:45 - 14:10 h, Room: H 3503, Talk 2

Genetha Anne Gray
Calculating and using sensitivity information during derivative-free optimization routines

Coauthors: Ethan Chan, John Guenther, Herbie Lee, John Siirola


The incorporation of uncertainty quantification (UQ) into optimization routines can help identify, characterize, reduce, and possibly eliminate uncertainty while drastically improving the usefulness of computational models and optimal solutions. Current approaches are in that they first identify optimal solutions and then, perform a series of UQ runs using these solutions. Although this approach can be effective, it can be computationally expensive or produce incomplete results. Model analysis that takes advantage of intermediate optimization iterates can reduce the expense, but the sampling done by the optimization algorithms is not ideal. In this talk, we discuss a simultaneous optimization and UQ approach that combines Bayesian statistical models and derivative-free optimization in order to monitor and use sensitivity information throughout the algorithm's execution.



Tuesday, 14:15 - 14:40 h, Room: H 3503, Talk 3

Satyajith Amaran
A comparison of software and algorithms in unconstrained simulation optimization problems

Coauthors: Scott J. Bury, Nikolaos V. Sahinidis, Bikram Sharda


Over the last few decades, several algorithms for simulation optimization (SO) have appeared and, along with them, diverse application areas for these algorithms. The algorithmic approaches proposed in the literature include ranking and selection, sample average approximation, metaheuristics, response surface methodology and random search. Application areas range from urban traffic control to investment portfolio optimization to operation scheduling. However, a systematic
comparison of algorithmic approaches for simulation optimization problems from the literature is not available. At this juncture in the evolution of SO, it is instructive to review the size and kinds of problems handled as well as the performance of different classes of algorithms, both in terms of
quality of solutions and number of experiments (or function evaluations) required. In this work, we use a library of diverse algorithms, and propose a method to assess their performance under homogeneous and heterogeneous variances on a recently-compiled simulation optimization test set. Discussions follow.


  cash loans . Therefore, we can say that the active substances in its composition are more perfectly mixed. Vardenafil is not only present in the original Cheap Levitra, but also as part of its analogs.