Automated Design: A Journey Across Modelling, Optimization, and Education

Machine intelligence greatly impacts almost all domains of our societies. It is profoundly changing the field of mechanical engineering with new technical possibilities and processes. The education of future engineers also needs to adapt in terms of techniques and even skills.

Using the design of electro-mechanical actuators as a common thread, this work explores the many-facets of automated design: modeling, optimization, and education, and looks for the prerequisites essential to its successful application.

The journey starts by building a modular and integrated model. It focuses on the prediction of system-level specifications that yield high added-value for decision-makers and shorten the path from the model to the final product. Combined with multi-objective evolutionary algorithms (MOEAs) and visualization tools, the model forms an automated design tool that helps engineers and decision-makers to rapidly get important insights into their design task. Its potential and benefits are validated through two specific applications. The results, however, also highlight a gap between the reported performance of optimizers on common benchmark problems and the actual performance on these problems.

To further develop optimizers, appropriate and realistic benchmark problems are needed. A subset of the integrated design model is used to formulate a new test suite called MODAct, composed of 20 constrained multi-objective optimization problems (CMOPs) with variable levels of complexity. In addition, numerical approaches to evaluate the constraint landscape of CMOPs are introduced and applied to identify the differences in features of MODAct against 45 benchmark problems from literature. Further, the convergence performance of three algorithms on the same problems highlights the key role of constraints and, in particular, the number of simultaneously violated constraints in MODAct problems.

In a next step, existing constraint handling strategies suitable for MOEAs along with a newly proposed technique for many-constraint problems are evaluated. Their parameters are tuned for different problems. The performance of the various configurations further highlights the difference between MODAct and other benchmark problems and show the highly competitive results of the proposed constraint handling technique on realistic design problems.

As the technical limits are removed, the impact of automated design on the work of future engineers should be considered. On the one hand, the development of professional skills by students working on team project in different settings has been evaluated thanks to 205 students from three classes. Explicitly addressing these skills within the project seems key to support stronger and broader learning, suggesting changes that do not require a full curriculum redesign. On the other hand, nine groups (33 students) have been asked to design an actuator using a conventional approach followed by an automated design approach. The actuators suggested by students using the automated tool outperform the designs obtained through the traditional approach. Six groups even suggest solutions cheaper, three of which are also smaller, than the product of experienced industry engineers. Students proved thus capable of leveraging the tool within a short time. The analysis of their mistakes suggests possible improvements for future tools. As these students leave university, they carry the hope to see such methods spread in industry.

Multi-objective optimisation applied to industrial energy problems

Geoffrey Basil Leyland

This thesis presents the development of a new multi-objective optimisation tool and applies it to a number of industrial problems related to optimising energy systems. Multi-objective optimisation techniques provide the information needed for detailed analyses of design trade-offs between conflicting objectives. For example, if a product must be both inexpensive and high quality, the multi-objective optimiser will provide a range of optimal options from the cheapest (but lowest quality) alternative to the highest quality (but most expensive), and a range of designs in between – those that are the most interesting to the decision-maker. The optimisation tool developed is the queueing multi-objective optimiser (QMOO), an evolutionary algorithm (EA). EAs are particularly suited to multi-objective optimisation because they work with a population of potential solutions, each representing a different trade-off between objectives. EAs are ideal to energy system optimisation because problems from that domain are often non-linear, discontinuous, disjoint, and multi-modal. These features make energy system optimisation problems difficult to resolve with other optimisation techniques. QMOO has several features that improve its performance on energy systems problems – features that are applicable to a wide range of optimisation problems. QMOO uses cluster analysis techniques to identify separate local optima simultaneously. This technique preserves diversity and helps convergence to difficult-to-find optima. Once normal dominance relations no longer discriminate sufficiently between population members certain individuals are chosen and removed from the population. Careful choice of the individuals to be removed ensures that convergence continues throughout the optimisation. Preserving of the "tail regions" of the population helps the algorithm to explore the full extent of the problem's optimal regions. QMOO is applied to a number of problems: coke factory placement in Shanxi Province, China; choice of heat recovery system operating temperatures; design of heat-exchanger networks; hybrid vehicle configuration; district heating network design, and others. Several of the problems were optimised previously using single-objective EAs. QMOO proved capable of finding entire ranges of solutions faster than the earlier methods found a single solution. In most cases, QMOO successfully optimises the problems without requiring any specific tuning to each problem. QMOO is also tested on a number of test problems found in the literature. QMOO's techniques for improving convergence prove effective on these problems, and its non-tuned performance is excellent compared to other algorithms found in the literature.

EPFL2002

Impacts of Constraints and Constraint Handling Strategies for Multi-Objective Mechanical Design Problems

Cyril Picard, Jürg Alexander Schiffmann

Multi-objective optimization tools are becoming increasingly popu-lar in mechanical engineering and allow decision-makers to better understand the inevitable trade-offs. Mechanical design problems can however combine properties that make the use of optimization more complex: (i) expensive cost functions; (ii) discrete or step-like behavior of the cost functions; and (iii) non-linear constraints. The latter in particular has a great impact on the convergence and the diversity of the obtained Pareto front. In this paper, we present five bi-objective mechanical design optimization problems with various levels of constraint complexity. They are used to rigorously benchmark two common constraint handling strategies (constrained-dominance and penalty function). The results suggest that both strategies have similar performance, and that as constraints become more intricate, convergence to the best-known Pareto front is not guaranteed. Indeed, analyzing the evolution of the hypervolume along generations reveals that the optimizer can get trapped in local optima. A detailed analysis of the obtained Pareto fronts for the proposed problems allows us to qualify the effects of the different constraints.

2018

Robust Adaptive Decision Making: Bayesian Optimization and Beyond

Ilija Bogunovic

The central task in many interactive machine learning systems can be formalized as the sequential optimization of a black-box function. Bayesian optimization (BO) is a powerful model-based framework for \emph{adaptive} experimentation, where the primary goal is the optimization of the black-box function via sequentially chosen decisions. In many real-world tasks, it is essential for the decisions to be \emph{robust} against, e.g., adversarial failures and perturbations, dynamic and time-varying phenomena, a mismatch between simulations and reality, etc. Under such requirements, the standard methods and BO algorithms become inadequate. In this dissertation, we consider four research directions with the goal of enhancing robust and adaptive decision making in BO and associated problems. First, we study the related problem of level-set estimation (LSE) with Gaussian Processes (GPs). While in BO the goal is to find a maximizer of the unknown function, in LSE one seeks to find all "sufficiently good" solutions. We propose an efficient confidence-bound based algorithm that treats BO and LSE in a unified fashion. It is effective in settings that are non-trivial to incorporate into existing algorithms, including cases with pointwise costs, heteroscedastic noise, and multi-fidelity setting. Our main result is a general regret guarantee that covers these aspects. Next, we consider GP optimization with robustness requirement: An adversary may perturb the returned design, and so we seek to find a robust maximizer in the case this occurs. This requirement is motivated by, e.g., settings where the functions during optimization and implementation stages are different. We propose a novel robust confidence-bound based algorithm. The rigorous regret guarantees for this algorithm are established and complemented with an algorithm-independent lower bound. We experimentally demonstrate that our robust approach consistently succeeds in finding a robust maximizer while standard BO methods fail. We then investigate the problem of GP optimization in which the reward function varies with time. The setting is motivated by many practical applications in which the function to be optimized is not static. We model the unknown reward function via a GP whose evolution obeys a simple Markov model. Two confidence-bound based algorithms with the ability to "forget" about old data are proposed. We obtain regret bounds for these algorithms that jointly depend on the time horizon and the rate at which the function varies. Finally, we consider the maximization of a set function subject to a cardinality constraint

k

in the case a number of items

\tau

from the returned set may be removed. One notable application is in batch BO where we need to select experiments to run, but some of them can fail. Our focus is on the worst-case adversarial setting, and we consider both \emph{submodular} (i.e., satisfies a natural notion of diminishing returns) and \emph{non-submodular} objectives. We propose robust algorithms that achieve constant-factor approximation guarantees. In the submodular case, the result on the maximum number of allowed removals is improved to

\tau = o(k)

in comparison to the previously known

\tau=o(\sqrt{k})

. In the non-submodular case, we obtain new guarantees in the support selection and batch BO tasks. We empirically demonstrate the robust performance of our algorithms in these, as well as, in data summarization and influence maximization tasks.

EPFL2019