Offshore wind auction simulation

该EngD的目的是开发一种以新型建模方法为基础的工具，该方法可供战略团队在一系列场景和输入下模拟CfD拍卖。该模型的输出将使人们能够研究拍卖的动态，从而确定最佳投标战略。该模型将附有一份支持文件，其中将包括每项战略的细节/定义，并解释各种建模假设。这一压倒一切的目标可以通过完成三个主要目标来实现。第一个目标是利用EDF的海上风电成本分析工具（OWCAT）评估主要竞争拍卖参与者的预期电价及其潜在收益的现有方法。由于影响模型输入的不确定性来源多种多样，因此需要建立随机成本模型。这将被用来创建一个概率分布的LCOE估计的每个主要项目竞争在CfD拍卖。第二个目标是开发一种方法，可以模拟竞争对手的行为在拍卖。不同的竞争者可以通过评估他们可能的行为（例如，考虑文化差异如何影响投标策略）分为不同的“拍卖类型”。该方法将利用博弈论和拍卖理论来模拟拍卖动态。通过前两个目标实现的工作将结合起来，以开发一个综合模型，该模型可以模拟各种不同情况下的拍卖。该模型将利用蒙特卡罗模拟运行拍卖多次，使用竞争对手预测的投标价格的概率分布作为每次模拟中的不同数据点。最终目标是将创建的模型实施到实际项目中，以帮助EDF做出决策。

The aim of this EngD is to develop a tool underpinned by a novel modelling approach which can be used by strategy teams to simulate CfD auctions under a range of scenarios and inputs. The outputs from the model will allow for the dynamics of the auction to be studied, allowing for an optimum bidding strategy to be determined. The model will be accompanied by a supporting document which will include details / definitions of each strategy and explain the various modelling assumptions. This overriding aim can be achieved through completion of three main objectives.The first objective is to build on existing methodologies which utilise EDF's Offshore Wind Cost Analysis Tool (OWCAT) to assess the expected electricity prices of major competing auction participants and their potential yields. Since there are a variety of sources of uncertainty affecting the model inputs, stochastic cost modelling is required. This will be used to create a probability distribution of LCOE estimates for each of the major projects competing at CfD auction.The second objective is to develop a methodology which can simulate a competitor's behaviour at auction. Separate competitors can be categorised into different 'auction types' through assessing their likely behaviour (e.g. considering how cultural differences may affect bidding strategies). The methodology will make use of game theory and auction theory to simulate auction dynamics.Work achieved through the first two objectives will be combined to develop a comprehensive model, which can model the auction under a range of different scenarios. The model will make use of Monte Carlo simulation to run the auction many times, using the probability distribution of competitors predicted bid price as different data points in each simulation. The final objective is to implement the created model into a live project to assist in decision making at EDF.