Martin H. Krieger

Energy-delay tradeoffs in smartphone applications

Many applications are enabled by the ability to capture videos on a smartphone and to have these videos uploaded to an Internet-connected server. This capability requires the transfer of large volumes of data from the phone to the infrastructure. Smartphones have multiple wireless interfaces -- 3G/EDGE and WiFi -- for data transfer, but there is considerable variability in the availability and achievable data transfer rate for these networks. Moreover, the energy costs for transmitting a given amount of data on these wireless interfaces can differ by an order of magnitude. On the other hand, many of these applications are often naturally delay-tolerant, so that it is possible to delay data transfers until a lower-energy WiFi connection becomes available. In this paper, we present a principled approach for designing an optimal online algorithm for this energy-delay tradeoff using the Lyapunov optimization framework. Our algorithm, called SALSA, can automatically adapt to channel conditions and requires only local information to decide whether and when to defer a transmission. We evaluate SALSA using real-world traces as well as experiments using a prototype implementation on a modern smartphone. Our results show that SALSA can be tuned to achieve a broad spectrum of energy-delay tradeoffs, is closer to an empirically-determined optimal than any of the alternatives we compare it to, and, can save 10-40% of battery capacity for some workloads.

Six propositions on the poor and pollution

The effect of improving the environment may be greater inequities in our society. Current environmental programs maintain this inequity, proposed environmental programs may make things worse, and even if we do improve the environment, contentment may decrease. A political coalition of environmental and equity enthusiasts may provide a viable way out of these dilemmas.

Commentary: Pervasive Urban Media Documentation

Jeongyeup Paek is a doctoral student of computer science in the Viterbi School of Engineering at the University of Southern California. In earlier commentaries, we have described efforts to document urban life, using still photography and aural recording (Krieger 2004; Krieger and Holman 2007). Characteristic of these efforts is the desire to be systematic and comprehensive, covering a very wide range of examples if not a complete inventory. And they have been concerned not so much with individual sites or places as with the choreography of urban life and the industrial engineering that enables large numbers of people to live and thrive nearby each other. So an archive has been produced with suites of images or aural recordings based on location, topic, or similarly functioning buildings. Tufte (1990) has called such efforts “small multiples,” albeit in the context of information graphics. Diderot so cataloged the industrial processes of his day, Charles Marville photographed the streets of Paris before and after they were eviscerated and reconstructed by Napoleon III (for the latter, we rephotographed those sites in 2008; see http://www.usc.edu/sppd/parismarville). Crucially, all such efforts require that images or recordings be associated with place, time, or list in a catalog. That is, indexing and mapping turns a collection of records into an archive. In the cinematic arts and in medical and engineering imaging, there has developed a notion of montage, showing the world from various aspects. Motion pictures employ cuts and montage to convey a multiaspectival view of the world and narrative, with flashbacks, point-of-view shots, and various degrees of wide-angle and narrowangle shots, all interdigitated to give a richer sense of the world in the service of storytelling. We have learned to make presentations that allow people to integrate this presentation, a movie, into a sense of a continuous whole world. The novel and the epic are, of course, precursors. Tomography shows slices of body organs, so allowing for a mental, more three-dimensional reconstruction by trained physicians. Phenomenologically, we are able to take a multiaspectival presentation and use it to revise our initial intuitions, and so fill in details, into a rich complex story or world. Sometimes, we need to abandon our initial intuitions, and then we try again to see the world as whole. In the cinematic arts, these processes are characterized as mastering a “screen language,” and in phenomenology we speak of “fulfillment” and “unity in multiplicity” or “identity in manifolds” (Daley 2003; Sokolowski 2000). We can have such pervasive, dense, multiaspectival documentation of an urban place. GPS-tagged photos are now readily made, and they are stitched together into very large total panoramas. Google Street View is perhaps the epitome of such an effort, and there are extensions that allow for three-dimensional reconstructions. Qik Commentary: Pervasive Urban Media Documentation

Modeling urban change

Abstract In this paper the simulation of crisis phenomena in complex social systems is explored. Analogies from complex physical systems containing a “large” number of objects are used to guide the modeling effort. A list processing program is developed that simulates the changing racial balance in urban areas.

Theorems as meaningful cultural artifacts: Making the world additive

Mathematical theorems are cultural artifacts and may be interpreted much as works of art, literature, and tool-and-craft are interpreted. The Fundamental Theorem of the Calculus, the Central Limit Theorem of Statistics, and the Statistical Continuum Limit of field theories, all show how the world may be put together through the “arithmetic addition” of suitably prescribed parts (velocities, variances, and renormalizations and scaled blocks, respectively). In the limit — of smoothness, statistical independence, and large N — higher-order parts, such as accelerations, are, for the most, part irrelevant, affirming that, in the end, most of the worlds particulars may be averaged over (a very un-Scriptural point of view). (We work out all of this in technical detail, including a nice geometric picture of stochastic integration, and a method of calculating the variance of the sum of dependent random variables using renormalization group ideas.) These fundamental theorems affirm a culture that is additive, ahistorical, Cartesian, and continuist, sharing in what might be called a species of modern culture. We understand mathematical results as useful because, like many other such artifacts, they have been adapted to fit the world, and the world has been adapted to fit their capacities. Such cultural interpretation is in effect motivation for the mathematics, and might well be offered to students as a way of helping them understand what is going on at the blackboard. Philosophy of mathematics might want to pay more attention to the history and detailed technical features of sophisticated mathematics, as a balance to the usual concerns that arise in formalist or even Platonist positions.

The physicist’s toolkit

A physicist’s ‘‘toolkit’’ might include mathematical, diagrammatic, and modeling tools, including such models as a crystalline solid, a harmonic oscillator, or an inverse square law of interaction. The notion of a toolkit provides a meeting ground for scientists, philosophers, and teachers for appreciating what scientists do. For doing science may be thought of as a craft, skillfully employing a kit of tools.

Contingency in Planning: Statistics, Fortune, and History

action skills are fully developed. Above all, foreign Ph.D. students cannot be seen as temporary visitors who will go back to teach in some far off land. They absorb a sizable portion of our faculty resources and they represent the next generation of planning educators, not only abroad but in North America as well. Many of them are already teaching our masters students while working on their Ph.D., and many will stay and teach on a short or long term basis. It is good to see that the new ACSP commission on Ph.D. education is starting to talk about these issues. Hopefully, they will find the resources to conduct thorough evaluations and develop appropriate support programs that will give foreign students that best possible education. As our Ph.D. education goes, so goes the next generation of planning professionals they will educate. We cannot afford to let foreign Ph.D. students slip by with narrow, myopic educations any more than we can afford to let this happen to domestic Ph.D. students. Leonard F. Heumann

Where do centers come from

A ‘center’ is a marked place in space or time or in a collection of objects, and surrounding it is a structure or pattern that supports it. The questions that concern us here are: ‘Why are there centers at all?’ ‘Why is a center at a particular X?’ Historical and combinatorial processes of centralization are reviewed, and a phenomenology and mechanisms are provided. Models considered include: stochastic markets with increasing returns to scale, codes as in DNA, combinatoric processes as in statistical mechanics, and differentiation in biology. Polycenters.

Taking Pictures in the City

Planners, and teachers and students of planning, ought systematically to document ordinary urban phenomena: store-front churches, vernacular visual merchandising, industry,... Careful archiving and indexing is crucial. For long-lasting records, photographic film still provides the best medium for still photographs; digital video and audio are probably best for motion pictures and sound.

Segmentation and Filtering into Neighborhoods as Processes of Percolation and Diffusion: Stochastic Processes (Randomness) as the Null Hypothesis

A set of models is reviewed in which orderly urban phenomena are accounted for in terms of noise (randomness, stochastic processes): segmentation as a process of runs in random processes (percolation); filtering as a process of random walk and so diffusion, and as a Markov process. In particular, ‘continuity’ of the spatial price envelope, a truism of real estate, leads to the diffusion equation. No foreordained centers, boundaries, or gravity (or entropy) is needed to obtain clustering in these models. The models are combined in a simulation, implemented in a Lotus 1-2-3 spreadsheet, as well as in a partial differential equation driven by a stochastic source term.