Optimization problems for controlled mechanical systems : bridging the gap between theory and application

Mechanical control systems have become a part of our everyday life. Systems such as automobiles, robot manipulators, mobile robots, satellites, buildings with active vibration controllers and air conditioning systems, make life easier and safer, as well as help us explore the world we live in and exploit it’s available resources. In this chapter, we examine a specific example of a mechanical control system; the Autonomous Underwater Vehicle (AUV). Our contribution to the advancement of AUV research is in the area of guidance and control. We present innovative techniques to design and implement control strategies that consider the optimization of time and/or energy consumption. Recent advances in robotics, control theory, portable energy sources and automation increase our ability to create more intelligent robots, and allows us to conduct more explorations by use of autonomous vehicles. This facilitates access to higher risk areas, longer time underwater, and more efficient exploration as compared to human occupied vehicles. The use of underwater vehicles is expanding in every area of ocean science. Such vehicles are used by oceanographers, archaeologists, geologists, ocean engineers, and many others. These vehicles are designed to be agile, versatile and robust, and thus, their usage has gone from novelty to necessity for any ocean expedition.

Tackling ‘Wicked’ problems holistically with institutionalist policymaking

One of our most pressing needs in creating a more sustainable world is the explicit development of holistic policy. This is becoming increasingly apparent as we are faced with more and more ‘wicked problems', the most difficult class of problems that we can conceptualize. Such problems consist of ‘clusters’ of problems, and include socio-political and moral-spiritual issues. This paper articulates a methodology that can be applied to the analysis and design of underlying organizational structures and processes that will consistently and effectively address wicked problems while being consistent with the advocated 'learning by doing' approach to change management and policy making. This transdisciplinary methodology—known as the institutionalist policymaking framework—has been developed from the perspective of institutional economics synthesized with perspectives from ecological economics and system dynamics. In particular it draws on the work first presented in Hayden’s 1993 paper ‘Institutionalist Policymaking’—and further developed in his 2006 book, at the heart of which lies the SFM—and the applicability of this approach in tackling complex and wicked problems.

Gender and municipal politics : problems, perspectives and possibilities

Scholars of local government have repeatedly lamented the lack of literature on the subject (e.g., Mowbray 1997; Pini, Previte, Haslam & McKenzie 2007). As Dollery, Marshall and Worthington (2003: 1) have commented, local government has often been the ‘poor cousin of its more exalted relatives in terms of the attention it attracts from the research community.’ The exalted relatives Dollery et al. (2003) refer to are national political environments, where women’s participation has elicited significant attention. However, the dearth of research on the specific subject of women’s representation in local government is rarely acknowledged (Neyland & Tucker 1996; Whip & Fletcher 1999). This edited book attempts to redress this situation. Each chapter applies an explicit gender analysis to their specific topic of focus, making ‘gender visible in social phenomenon; [and] asking if, how, and why social processes, standards, and opportunities differ systematically for women and men’ (Howard, Risman & Sprague 2003: 1). These analyses in the local government context are critical for understanding the extent and nature of balanced representation at all levels of government. Furthermore, some women start their elective careers serving on school boards, city or town councils or as mayors, before progressing to state and national legislative offices. Hence, the experiences of women in local government illustrate broader notions of democracy and may for some individual women, shape their opportunities further along the political pipeline.

Numerical study of two ill-posed one-phase Stefan problems

We treat two related moving boundary problems. The first is the ill-posed Stefan problem for melting a superheated solid in one Cartesian coordinate. Mathematically, this is the same problem as that for freezing a supercooled liquid, with applications to crystal growth. By applying a front-fixing technique with finite differences, we reproduce existing numerical results in the literature, concentrating on solutions that break down in finite time. This sort of finite-time blow-up is characterised by the speed of the moving boundary becoming unbounded in the blow-up limit. The second problem, which is an extension of the first, is proposed to simulate aspects of a particular two-phase Stefan problem with surface tension. We study this novel moving boundary problem numerically, and provide results that support the hypothesis that it exhibits a similar type of finite-time blow-up as the more complicated two-phase problem. The results are unusual in the sense that it appears the addition of surface tension transforms a well-posed problem into an ill-posed one.

A Stochastic View of Optimal Regret through Minimax Duality

We study the regret of optimal strategies for online convex optimization games. Using von Neumann's minimax theorem, we show that the optimal regret in this adversarial setting is closely related to the behavior of the empirical minimization algorithm in a stochastic process setting: it is equal to the maximum, over joint distributions of the adversary's action sequence, of the difference between a sum of minimal expected losses and the minimal empirical loss. We show that the optimal regret has a natural geometric interpretation, since it can be viewed as the gap in Jensen's inequality for a concave functional--the minimizer over the player's actions of expected loss--defined on a set of probability distributions. We use this expression to obtain upper and lower bounds on the regret of an optimal strategy for a variety of online learning problems. Our method provides upper bounds without the need to construct a learning algorithm; the lower bounds provide explicit optimal strategies for the adversary. Peter L. Bartlett, Alexander Rakhlin

The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network

Sample complexity results from computational learning theory, when applied to neural network learning for pattern classification problems, suggest that for good generalization performance the number of training examples should grow at least linearly with the number of adjustable parameters in the network. Results in this paper show that if a large neural network is used for a pattern classification problem and the learning algorithm finds a network with small weights that has small squared error on the training patterns, then the generalization performance depends on the size of the weights rather than the number of weights. For example, consider a two-layer feedforward network of sigmoid units, in which the sum of the magnitudes of the weights associated with each unit is bounded by A and the input dimension is n. We show that the misclassification probability is no more than a certain error estimate (that is related to squared error on the training set) plus A3 √((log n)/m) (ignoring log A and log m factors), where m is the number of training patterns. This may explain the generalization performance of neural networks, particularly when the number of training examples is considerably smaller than the number of weights. It also supports heuristics (such as weight decay and early stopping) that attempt to keep the weights small during training. The proof techniques appear to be useful for the analysis of other pattern classifiers: when the input domain is a totally bounded metric space, we use the same approach to give upper bounds on misclassification probability for classifiers with decision boundaries that are far from the training examples.

Open problems in the security of learning

Machine learning has become a valuable tool for detecting and preventing malicious activity. However, as more applications employ machine learning techniques in adversarial decision-making situations, increasingly powerful attacks become possible against machine learning systems. In this paper, we present three broad research directions towards the end of developing truly secure learning. First, we suggest that finding bounds on adversarial influence is important to understand the limits of what an attacker can and cannot do to a learning system. Second, we investigate the value of adversarial capabilities-the success of an attack depends largely on what types of information and influence the attacker has. Finally, we propose directions in technologies for secure learning and suggest lines of investigation into secure techniques for learning in adversarial environments. We intend this paper to foster discussion about the security of machine learning, and we believe that the research directions we propose represent the most important directions to pursue in the quest for secure learning.

Overcoming the complexity of diagnostic problems due to sensor network architecture

In fault detection and diagnostics, limitations coming from the sensor network architecture are one of the main challenges in evaluating a system’s health status. Usually the design of the sensor network architecture is not solely based on diagnostic purposes, other factors like controls, financial constraints, and practical limitations are also involved. As a result, it quite common to have one sensor (or one set of sensors) monitoring the behaviour of two or more components. This can significantly extend the complexity of diagnostic problems. In this paper a systematic approach is presented to deal with such complexities. It is shown how the problem can be formulated as a Bayesian network based diagnostic mechanism with latent variables. The developed approach is also applied to the problem of fault diagnosis in HVAC systems, an application area with considerable modeling and measurement constraints.

Floods, earthquakes and insurance coverage : issues, problems and solutions

The recent floods in Queensland and elsewhere in Australia as well as the recent earthquakes in New Zealand have again given rise to very significant uninsured losses. This article looks at the issue of cover protection against catastrophes such as floods and earthquakes affecting home buildings and contents insurance and the standard cover provisions of the Insurance Contracts Act 1984 (Cth). It points also to the possibility of a national scheme to cover natural disasters including floods.