Algorithms for load balancing in electricity markets and data centers

of the Dissertation Algorithms For Load Balancing In Electricity Markets And Data Centers by Bochao Shen Doctor of Philosophy in Computer Science Northeastern University, April 2018 Dr. Ravi Sundaram, Adviser Electricity and computers are two corner stones of this information era. Energy and computation are critical resources in perennial short supply because our consumption continues to grow day by day. Increasing supply in a substantial way requires fundamental advances in energy generation and computing technology, but such advances are few and far between. Load balancing is an important technique for mitigating the issue of scarce resources. We broadly interpret load balancing to include both the optimization of resource distribution as well as the management of end-user demand. In this thesis, we study algorithms for load balancing in electricity markets and data centers. First, we study the temporal load balancing problem in electricity markets where peak demand and supply-demand imbalance are major problems. It is often suggested that exposing consumers to real-time pricing will incentivize them to change their usage and mitigate the problem. However, we show that risk-averse electricity consumers react to price fluctuations by scaling back on their total demand, leading to the unintended consequence of an overall decrease in production/consumption and reduced economic efficiency. Compared with the relatively fixed production mode of electricity power (the supply), the consumption pattern of end users (the demand) is more variable and potentially changeable. This makes temporally shifting consumers’ electricity load possible. We propose SmartShift, a new scheme that allows households to move their demands from peak hours in exchange for greater electricity consumption in non-peak hours. We show that our scheme not only enables increased consumption and consumer welfare but also allows the distribution company to increase profits. We next consider the fault-tolerant spatial load balancing problem in data centers where computational loads get balanced by being assigned on different locations (machines computational resources). k-HA (high-availability), a fault tolerance property of virtual machine (VM) placement