Tutorial Sessions

Technology for sensing and actuation at microscopic length scales has advanced rapidly in recent decades, enabling time and spatially resolved measurements at micrometer and even nanometer scales.  Modeling of the dynamics of a collection of discrete particles has also become more accurate, enabled by advances in measurement and computational technologies.  High-fidelity models of the forces acting on or between particles are now available for many technologically relevant systems.  For many cases of practical interest, stochastic simulations must be performed to predict the dynamic evolution of the many-body system, based on the force models.  When a simpler approximation for ensemble-averaged system states is available, it generally takes the form of a nonlinear stochastic differential equation, such as a Langevin equation.  The best way to construct such a coarse-grained equation is still an open research question.

There is great practical interest in being able to produce materials with predefined structures with microscopic dimensions.  For example, if nanoparticles are arranged into a perfect crystal, they exhibit new optical properties that could enable light-based versions of electronics, and even futuristic concepts such as cloaking (invisibility) devices.  However, extremely low defect density is required to achieve these novel properties.  Mechanical manipulation can be used to place individual particles (“top-down”), while fields such as temperature and electric field can be used to control much larger ensembles of particles (“bottom-up”).  Bottom-up directed assembly may be more scalable to high-throughput manufacturing, but it is also inherently underactuated.  Feedback control is typically not used during directed self-assembly, but provides a potential solution to defect control in nanomaterials assembly.  Examples, challenges, and possibilities for controlling a directed self-assembly processes will be reviewed in the context of the available and emerging technologies for modeling, sensing and actuation. 

Bevan has developed direct, nonintrusive optical microscopy methods and inverse analyses to measure and model interactions, dynamics, and structures in many-body colloidal systems with a focus on self-assembly problems.  Shapiro studies feedback control of microscopic and nanoscopic objects by flow, electric field, and magnetic actuation.  Ford is an expert in statistical mechanics and has a long-standing collaboration with Bevan to develop equilibrium and dynamic models of colloidal systems from direct microscopic observations.  Grover applies dynamic modeling and model reduction to control of molecular ensembles at the micro- and nano-scales.

This session will highlight current needs and future challenges in control for the semiconductor industries. Applications will include both microelectronics and photovoltaics processing and manufacturing. The presenters include both industrial and academic researchers. The session is intended to communicate industrial needs to the ACC attendees, who are primarily academic researchers. Session will also be an opportunity for dialogue between the academic and industrial participants, with the goal of articulating and generating more general and theoretical problems that are motivated by these industrial needs.

The purpose of this tutorial session is to explain how control theoretic tools and associated mathematical concepts can be used in option trading. No previous knowledge of options will be assumed. After explaining the mechanics of options and introducing the requisite mathematical models, the speakers will present a number of examples to demonstrate application of various trading algorithms, option hedging techniques and the use of both technical and fundamental analysis. Finally, the session will include discussion of a number of new and exciting research problems for the control field. One main theme of this tutorial session is that trading concepts can be explained in the context of a basic feedback loop with the control corresponding to modulation of the amount invested as a function of time.

The target audience for this tutorial session is members of the control community that are seeking an easy-to-digest introduction to option trading from a systems theoretic point of view. In this context, our goal is to bring the attendee “up to speed” Accordingly, considerable time will be dedicated to tutorial material and existing literature in the financial journals. This tutorial material will also include a review of basic terminology such as margin, short selling, bid-ask spreads, liquidity, put and call options, volatility and various technical indicators.

Data centers are proliferating worldwide as the result of high demand from Information Technology services. To reduce the cost of data center operation, shared hosting platforms are becoming widely deployed so that computing capacity can be provisioned on-demand. Power consumption and heat management are becoming more important than before due to the increasing amount of power consumption of data centers, energy cost and environmental considerations.  However, it has been challenging to manage the shared IT resources in these environments due to complicated and usually time-varying resource demands of the workloads, application structures and interactions among applications sharing the platforms. Power management needs to operate at multiple time granularities, and address challenges posed by interactions among system components. Data center cooling control has to address the spatial and temporal non-uniformities due to the time-varying cooling demand and location dependent cooling efficiencies.

This tutorial introduces the data center operation management problem as an emerging industrial control application. The speakers are to provide an overview of data center operation management,  discuss a few typical applications of advance control and optimization techniques on workload and resource management, server power control, cooling management, and integrated workload, power and cooling control problems in data centers.  The talks are to cover the problem definitions, model development, control architectures and algorithms, and experimental evaluation results. The speakers are to discuss further opportunities for control applications in data centers as well. 

This session is part I of a two-session series aimed at exploring ways that the field of controls can help to address important societal challenges, specifically those associated with developing a sustainable infrastructure. The focus of this part is on transitioning to a sustainable energy portfolio. The session will start with a keynote speech followed by three short talks by the three panelists. The final part is dedicated to a panel discussion with audience participation. The speakers and panelists will be chosen to bring in diverse domain expertise as well as explore sources for interdisciplinary interactions. For more information and updates on the speakers/panelists, check the web site at  

http://www.cds.caltech.edu/controlsforsociety.

This session is part II of a two-session series aimed at exploring ways that the field of controls can help to address important societal challenges, specifically those associated with developing a sustainable infrastructure. The objective of this part is to increase awareness in the controls community about challenges associated with sustainable growth. The session will start with a keynote speech followed by three short talks by the three panelists. The final part is dedicated to a panel discussion with audience participation. The speakers and panelists will be chosen to bring in diverse domain expertise as well as explore sources for interdisciplinary interactions. For more information and updates on the speakers/panelists, check the web site at http://www.cds.caltech.edu/controlsforsociety.

The objective of this tutorial session is to define several representative mathematical problems that evolve from the basic needs of making a power system smart (or, self-automated), and to point out how different aspects of control and optimization theory can play an instrumental role in solving those needs. We will discuss power system controls as traditional design problems as well as in terms of a variety of other relevant topics such as system identification, state estimation, model reduction, and cyber-physical control. We will invite six speakers from top-ranked universities in the US who are currently taking an active interest in smart grid research. Their presentations will cover distinct problems while maintaining a thematic continuity. The session will be dedicated to Dr. Joe Chow on the occasion of his 60th birthday. Professor Chow is one of the most distinguished researcher and educator in the field of power systems and controls in the US, with a research career spanning more than 35 years.