Evolution Control for parallel ANN-assisted simulation-based optimization application to Tuberculosis Transmission Control

Abstract

In many optimal design searches, the function to optimise is a simulator that is computationally expensive. While current High Performance Computing (HPC) methods are not able to solve such problems efficiently, parallelism can be coupled with approximate models (surrogates or meta-models) that imitate the simulator in timely fashion to achieve better results. This combined approach reduces the number of simulations thanks to surrogate use whereas the remaining evaluations are handled by supercomputers. While the surrogates' ability to limit computational times is very attractive, integrating them into the overarching optimization process can be challenging. Indeed, it is critical to address the major trade-off between the quality (precision) and the efficiency (execution time) of the resolution. In this article, we investigate Evolution Controls (ECs) which are strategies that define the alternation between the simulator and the surrogate within the optimization process. We propose a new EC based on the prediction uncertainty obtained from Monte Carlo Dropout (MCDropout), a technique originally dedicated to quantifying uncertainty in deep learning. Investigations of such uncertainty-aware ECs remain uncommon in surrogateassisted evolutionary optimization. In addition, we use parallel computing in a complementary way to address the high computational burden. Our new strategy is implemented in the context of a pioneering application to Tuberculosis Transmission Control. The reported results show that the MCDropout-based EC coupled with massively parallel computing outperforms strategies previously proposed in the field of surrogate-assisted optimization.

Parallel surrogate-assisted optimization: Batched Bayesian Neural Network-assisted GA versus q-EGO

“…In [39], we proposed a BNN-assisted GA and applied it successfully on a TBTC problem. The Evolution Controls implemented either focused on exploitation as the minimization of the predicted cost or on exploration as the maximization of the MCDropout-based BNN uncertainty and the maximization of the distance to the already simulated solutions.…”

“…Here, on each CPU core, a local search is run and its resulting candidate solution is simulated. In [39], the batch of solutions to simulate, produced according to the Evolution Control, is simulated in parallel.…”