Stochastic programming-approaches for integrating operational and financial decisions

该研究的目的是开发数学规划模型和解决方法的重要问题，在生产和物流，整合运营和财务的角度。随着市场的日益全球化和行业的重组，例如电力行业，出现了新的机会，为建模问题，运营和财务决策需要共同考虑。此外，市场和产业的日益自由化，使现代社会重要资源的价格和数量的不确定性空前增加。传统模型将财务和运营决策分开并独立处理。这些典型模型的决策在现代环境中可能非常不相关。为此，我们的目标是开发独特的随机规划模型的问题，在能源和交通的金融工具/机制或真实的选项的决定是普遍的。将开发基于数学规划分解和遗传学的解决方法，利用模型的特定结构。我们开发了两类随机模型。第一类问题将集中在基于优化的框架，以评估能源合同和其他衍生品/合同的资产组合，这些资产很难在标准方法下评估。另一类模型和方法将侧重于不确定情况下的运输采购，其中使用（组合）拍卖等金融框架将业务分配给相互竞争的运输供应商。这些新的模型及其解决方案将大大改善社会许多重要领域的决策支持，如电力和交通部门。

The objective of the research is to develop mathematical programming models and solution methods for important problems in production and logistics that integrate both operational and financial perspectives. With increasing globalization of markets and re-structuring of industries e.g. power industry new opportunities arise for modeling problems where operational and financial decisions need to be jointly considered. Furthermore, the increasing liberalization of markets and industries has created an unprecedented increase in uncertainty in prices and quantities for vital resources in modern society. Traditional models treat the financial and operational decisions separately and independently. Decisions from these typical models can be vastly irrelevant in modern environments. To this end, we aim to develop unique stochastic programming models for problems in energy and transportation where financial instruments/mechanisms or real option decisions are prevalent. Solution methods based on mathematical programming decomposition and heuristics will be developed that exploit the specific structure of the models. We develop two classes of stochastic models. The first class of problems will focus on optimization-based frameworks to evaluate portfolios of energy contracts and other derivatives/contracts for assets that are difficult to evaluate under standard methods. Another class of models and methods will focus on transportation procurement under uncertainty where financial frameworks such as (combinatorial) auctions are used to allocate business to competing transportation providers. These new classes of models and their solution methods will enable greatly improved decision support in many vital areas of society such as the power and transportation sectors.