Colloquium
Fall 2014
All talks are from 3:454:45 p.m. in the Colloquium Room, unless otherwise specified.

Dec 03

Nov 20

Oct 24Hallucination Helps: Energy Efficient Virtual Circuit Routing Prof. Clifford Stein Columbia University Time: 12:00 PM
View Abstract
In this talk, we give online and offline approximation algorithms for energy efficient circuit routing protocols for a network of components that are speed scalable, and that may be shutdown when idle. We will consider both the case where the speedscalable components are edges and the case when they are nodes. For the case of edges, we describe a polynomialtime offline algorithm that has a polylog approximation ratio. The key step of the algorithm design is a random sampling technique that we call hallucination. The algorithm extends rather naturally to an online algorithm. The analysis of the online algorithm introduces a natural ``priority'' multicommodity flow problem, and bounds the priority multicommodity flowcut gap. We will then explain the additional difficulty faced if the power management components are vertices, and explain how to how to surmount this difficulty in the offline case. This work spans several papers and is joint with Antonios Antoniadis, Nikhil Bansal, Sungjin Im, Ravishankar Krishnaswamy, Benjamin Moseley, Viswanath Nagarajan, Kirk Pruhs. 
Sep 24The USNA Powered Flight Program and a binary integer programming solution to its daily flight scheduling problem LT Peter Barkley USNA Time: 03:45 PM

Sep 18Change of phase for the humid atmosphere: a model of nonlinear discontinuous equation Prof. Roger Temam Indiana University Time: 12:00 PM
View Abstract
In this lecture we will recall the atmospheric equations of water vapor with saturation. In their simplest form, these equations form a nonlinear system of partial differential equations with discontinuities corresponding to the change of phase. We will address the issues of the definition of the solutions, in the context of convex analysis and variational inequalities, and establish some results on the existence, uniqueness and regularity of these solutions. 
Sep 16
View Abstract
Optimization is concerned with optimal performancefinding the best way to complete a task. It has been put to good use in a great number of diverse disciplines: advertising, agriculture, biology, business, economics, engineering, manufacturing, medicine, telecommunications, and transportation (to name but a few). In this lecture, we will showcase its amazing utility by demonstrating its applicability in the area of visual art, which at first glance would seem to have no use for it whatsoever! We will begin by describing how to use integer programming to construct a portrait out of complete sets of double nine dominoes. We will then describe how high quality solutions to certain largescale traveling salesman problems can lead to beautiful continuous line drawings. We will conclude by presenting other examples of Opt Artart constructed with the assistance of optimization techniques. 
Sep 10Optimal Patrol to Detect Attacks at Dispersed Heterogeneous Locations CDR Dick McGrath USNA Time: 03:45 PM
View Abstract
We study a patrol problem where several patrollers move between heterogeneous locations dispersed throughout an area of interest in order to detect enemy attacks. To formulate an effective patrol policy, the patrollers must take into account travel time between locations, as well as locationspecific parameters, which include patroller inspection times, enemy attack times, and cost incurred due to an undetected attack. We consider both random and strategic attackers. A random attacker chooses a location to attack according to a probability distribution, while a strategic attacker plays a twoperson zerosum game with the patrollers. In some cases, we can compute the optimal solution using linear programming. This method, however, becomes computationally intractable as the problem size grows. Therefore, we focus on developing efficient heuristics, based on aggregate index values, fictitious play, and shortest paths. Numerical experiments demonstrate that our heuristics produce excellent results with computation time orders of magnitude less than what is required to compute the optimal solution. 
Sep 03Optimal Future Sensor Allocation Problem Prof. Christopher Griffin Penn State University Time: 03:45 PM
View Abstract
In this talk we consider the problem of sensing a finite set of (moving) objects over a finite planning horizon using a set of sensors in prefixed locations that vary with respect to time over a discretized space. We formulate an integer programming problem that maximizes a quality of sensor return given either deterministic or probabilistic (i.e., forecasted) object routes. We examine the computational complexity of the problem and show it is NPHard. We analyze a representative data set of potential sensor collects and show, surprisingly, a near linear growth in computation time. We explain this discrepancy by identifying a class of subproblems (to which our data set belongs) that can be solved by a greedy heuristic. We then construct an artificial data set and illustrate nonlinear solution time growth. Finally, we define an alternate strongly polynomial (greedy) heuristic for solving the problem and investigate quality bounds. 
Aug 21Finite Time and Aperiodically Time Dependent Dynamics: The Research LandscapePast, Present, and Future Prof. Stephen Wiggins University of Bristol (UK) Time: 12:00 PM Abstract