Autonomous Electricity Trading Using Time-of-Use Tariffs in a Competitive Market
Authors: Daniel Urieli, Peter Stone
AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We extensively experiment with our algorithm in a large-scale, detailed electricity retail markets simulation of the Power Trading Agent Competition (Power TAC) and: 1) find that our algorithm results in 15% peak-demand reduction, 2) find that its peakflattening results in greater profit and/or profit-share for the broker and allows it to win against the 1st and 2nd place brokers from the Power TAC 2014 finals, and 3) analyze several economic implications of using TOU tariffs in competitive retail markets. [...] 5 Experimental Results We evaluated our TOU broker using paired tests. We measured the impact of modifying a component of the broker by testing the original and the modified version in a set of games, in which the opponents and most random factors in the simulation were held fixed (random seeds, weather conditions). |
| Researcher Affiliation | Academia | Daniel Urieli and Peter Stone Dept. of Computer Science The University of Texas at Austin Austin, TX 78712 USA {urieli,pstone}@cs.utexas.edu |
| Pseudocode | Yes | Algorithm 1 shows a pseudocode of our TOU optimization algorithm, which has two phases. |
| Open Source Code | No | The paper does not provide an explicit statement or link indicating that the source code for the described methodology is publicly available. |
| Open Datasets | No | Power TAC s simulator models real-world markets (like ERCOT s) with components of future markets, such as autonomous agents optimizing customer consumption. [...] Power TAC uses state-of-the-art customer models, which consume/produce using time-series generators based on real-world data. [...] To fix weather conditions, we used weather files containing 3 months of real-world weather. To cover year-round weather conditions we used 8 weather files (each file used by 1/8 of the games) with start-dates of January, April, July, October of 2009 and 2010. (No specific link, DOI, or formal citation with authors/year for public access to these real-world data or weather files is provided.) |
| Dataset Splits | No | The paper describes experiments run within the Power TAC simulation environment against other agents, including using "paired tests" and the "Wilcoxon matched-pairs signed-ranks test". However, it does not specify explicit training, validation, and test dataset splits in terms of percentages, sample counts, or references to predefined standard splits for reproducing data partitioning. |
| Hardware Specification | No | The paper describes the Power TAC simulation environment and its scale but does not provide any specific details about the hardware (CPU, GPU, memory, or cloud resources) used to run the experiments. |
| Software Dependencies | No | The paper mentions the Power TAC simulation environment and concepts like ERCOT, but it does not specify any software dependencies with version numbers (e.g., programming languages, libraries, frameworks, or specific simulation software versions) required for reproducibility. |
| Experiment Setup | Yes | Algorithm 1 shows a pseudocode of our TOU optimization algorithm, which has two phases. In a preliminary first phase (line 1) it uses the fixed-rate optimization algorithm of Urieli and Stone to get a reasonable fixed-rate seed for phase 2 (line 2-11). Phase 2 optimizes a TOU tariff using empirical gradient-ascent. First, the fixed-rate seed is converted into a TOU tariff with H identical entries (line 2). Next (lines 3-5), each of the entries is perturbed by ϵ and the utility UH(P) of the perturbed price vector is predicted using the predicted demand and cost curves (in our experiments ϵ = 0.5cents/k Wh). [...] We tested how using TOU tariffs optimized with Algorithm 1 affected 1) the broker s performance, and 2) the economy. We compared a TOU Broker using Algorithm 1 with two variations: one that uses fixed-rate tariffs and another that uses a naive TOU tariff optimization. We refer to these brokers as TOU, Fixed Rate, and TOUNaive. [...] We compared these three brokers in 2 different experiments, in which they played 200 games against (1) Agent UDE, and (2) CWIBroker, both of which use only fixed-rate tariffs. [...] To fix weather conditions, we used weather files containing 3 months of real-world weather. To cover year-round weather conditions we used 8 weather files (each file used by 1/8 of the games) with start-dates of January, April, July, October of 2009 and 2010. Opponents were fixed to be one of the top brokers played in the Power TAC 2014 finals: Agent UDE (Ozdemir and Unland 2015) (1st place), and CWIBroker (Hoogland and Poutre 2015) (2nd place). |