AI applications in future energy markets: Market implications and regulatory requirements

不断增加的间歇性可再生能源、分布式能源和消费者参与。因此，市场主体和系统运营商将面临一个更加复杂、动荡和不可预测的电力市场。与此同时，欧盟的透明度法规正在提供越来越多的公开可访问的市场和物理系统数据。算法交易和反向工程算法等人工智能(AI)工具为市场参与者和系统运营商提供了一条有希望的途径，以提高毛利率，减少交易努力，并增加更广泛的市场参与者的可及性。尽管有这些潜在的好处，但人们并不完全了解不断变化的市场动态的影响，在其他行业，类似的变化也导致了串通的结果。最近的一些串通算法定价的例子包括无意中的价格上涨，以及在亚马逊上创建和执行卡特尔。鉴于电力市场的独特性质，供求必须时刻匹配，操纵市场的可能性很大。尽管竞争和市场管理局(CMA)最近的调查强调了与这种潜在的串通行为相关的不确定性和风险，但这些都是限制因素。特别是，通过逆向工程深入了解竞争对手的投标策略，可以使代理人行使市场权力。反向工程其他市场参与者的交易策略也提出了关于人工智能道德及其对知识产权影响的重要问题。然而，反向工程这些投标策略也可以帮助检测串通行为。例如，纳斯达克证券交易所在2019年7月宣布，它将试验人工智能，以检测市场操纵和滥用的案件。可以想象，电力系统运营商和监管机构(例如Ofgem)可以从了解市场参与者的竞标做法中获益很大，以提高识别串通或操纵事件的可能性。为了理解这些对电力市场的影响，必须解决现有文献中的一个主要空白。在这种背景下，这个博士生计划回答以下问题：“关于能源系统的公开数据可以在多大程度上被用来对交易策略进行反向工程？”，“代理人在一个与今天的市场条件不同的市场中可能会有什么行为”，“拥有这样的信息会对市场的运作产生什么影响？”以及“需要如何制定监管以促进人工智能的应用”。为此，学生将研究用于能源市场应用的最先进的人工智能技术，开发一个多智能体评估框架，量化人工智能应用的市场影响，并为市场和监管框架的发展提供信息。

increasing amounts of intermittent renewables, distributed energy resources and consumer participation. As a result, market agents and system operators will face a more complex, volatile and unpredictable electricity market. Simultaneously, the EU's transparency regulations are providing a growing repository of publicly accessible market and physical system data. Artificial intelligence (AI) tools, such as algorithmic trading and reverse-engineering algorithms, present a promising avenue for market participants and system operators to increase gross margin, reduce trading efforts, and increase accessibility to a broader range of market participants. Despite these potential benefits, the impact of the changing market dynamics is not fully understood and in other industries similar changes have led to collusive outcomes. Some recent examples of collusive algorithmic pricing include inadvertent price rises, and creation and enforcement of cartels on Amazon. Given the unique nature of the electricity market where supply and demand must be matched at all times, the potential for market manipulation is great. Although recent investigations by the Competition and Market Authority (CMA) emphasize the uncertainty and risk associated with such potential collusive action as limiting factors. In particular, having in depth knowledge of competitors' bidding strategies through reverseengineering could allow an agent to exercise market power.Reverse-engineering other market actors' trading strategies also raises important questions regarding the ethics of AI and its implications for intellectual property rights.However, reverse-engineering these bidding strategies can also help detect collusive behaviour. For example, the NASDAQ stock exchange announced in July 2019 that it will be trialling AI to detect cases of market manipulation and abuse. It is conceivable that the electricity system operator and regulatory authorities (e.g. Ofgem) could benefit greatly from knowledge of market participants' bidding practices to improve the likelihood of discerning instances of collusion or manipulation. In order to understand these impacts on electricity markets, a major gap in the existing literature must be addressed.In this context, this PhD studentship proposes to answer the questions "to what extent publicly available data on the energy system can be used to reverse-engineer trading strategies?", "how agents may behave in a market with varying conditions to today's market", "what impact having such information would have on the functioning of the market?" and "how regulation needs to be developed to facilitate the applications of AI". In order to do so, the student will investigate the state-of-the-art AI techniques for energy market applications, develop a multi-agent assessment framework, quantify the market impact of AI applications and inform the development of market and regulatory framework.