Effects of Information and Time of Use Pricing on Irish Electricity Demand and Supply

In this dissertation I quantify residential behavior response to interventions designed to reduce electricity demand at different periods of the day. In the first chapter, I examine the effect of information provision coupled with bimonthly billing, monthly billing, and in-home displays, as well as a time-of-use (TOU) pricing scheme to measure consumption over each month of the Irish Consumer Behavior Trial. I find that time-of-use pricing with real time usage information reduces electricity usage up to 8.7 percent during peak times at the start of the trial but the effect decays over the first three months and after three months the in-home display group is indistinguishable from the monthly treatment group. Monthly and bi-monthly billing treatments are not found to be statistically different from another. These findings suggest that increasing billing reports to the monthly level may be more cost effective for electricity generators who wish to decrease expenses and consumption, rather than providing in-home displays. In the following chapter, I examine the response of residential households after exposure to time of use tariffs at different hours of the day. I find that these treatments reduce electricity consumption during peak hours by almost four percent, significantly lowering demand. Within the model, I find evidence of overall conservation in electricity used. In addition, weekday peak reductions appear to carry over to the weekend when peak pricing is not present, suggesting changes in consumer habit. The final chapter of my dissertation imposes a system wide time of use plan to analyze the potential reduction in carbon emissions from load shifting based on the Ireland and Northern Single Electricity Market. I find that CO2 emissions savings are highest during the winter months when load demand is highest and dirtier power plants are scheduled to meet peak demand. TOU pricing allows for shifting in usage from peak usage to off peak usage and this shift in load can be met with cleaner and cheaper generated electricity from imports, high efficiency gas units, and hydro units.

Learning Binary Code Representations for Effective and Efficient Image Retrieval

The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable problem for exhaustive similarity search algorithms. Hashing methods, which encodes high-dimensional descriptors into compact binary strings, have become very popular because of their high efficiency in search and storage capacity. In the first part, we propose a multimodal retrieval method based on latent feature models. The procedure consists of a nonparametric Bayesian framework for learning underlying semantically meaningful abstract features in a multimodal dataset, a probabilistic retrieval model that allows cross-modal queries and an extension model for relevance feedback. In the second part, we focus on supervised hashing with kernels. We describe a flexible hashing procedure that treats binary codes and pairwise semantic similarity as latent and observed variables, respectively, in a probabilistic model based on Gaussian processes for binary classification. We present a scalable inference algorithm with the sparse pseudo-input Gaussian process (SPGP) model and distributed computing. In the last part, we define an incremental hashing strategy for dynamic databases where new images are added to the databases frequently. The method is based on a two-stage classification framework using binary and multi-class SVMs. The proposed method also enforces balance in binary codes by an imbalance penalty to obtain higher quality binary codes. We learn hash functions by an efficient algorithm where the NP-hard problem of finding optimal binary codes is solved via cyclic coordinate descent and SVMs are trained in a parallelized incremental manner. For modifications like adding images from an unseen class, we propose an incremental procedure for effective and efficient updates to the previous hash functions. Experiments on three large-scale image datasets demonstrate that the incremental strategy is capable of efficiently updating hash functions to the same retrieval performance as hashing from scratch.