An important and constantly evolving area of research and application deals with predictive analytics for high-frequency and high-dimensional continuous-valued or discrete-valued time series, and merging the results from modeling such complex processes with decision/game theory methods for concluding ultimate risk/reward. Such analyses are increasingly important in two highly inter-disciplinary areas that bring together researchers and practitioners from computer science, engineers, finance, and statistics: analysis of financial systems and transportation systems. Another core context is that of multi-step forecasting– based on banks of formal statistical models of the dynamic, stochastic systems that drive financial and economic markets– in which new kinds of multi-step ahead utility functions to reflect core costs/benefits of decision makers are increasingly sought. While dynamic programming ideas are old and well understood, implementing specific optimized policies based on a complex model/multi-step utility construct remains a core challenge in anything but rather stylized contexts. Multiple new ideas have recently emerged to address this need to bring together the theory of rational behaviors in the face of uncertainty with the real-world practicalities of increasingly model and utility function complexity. Among these, methods of Bayesian model
emulation, defining surrogate models of complex dynamic systems coupled with the optimization requirements to explore and quantitatively understand consequences of actions, have become visible in areas such as portfolio analysis in financial systems. The basic ideas underlying such approaches are novel and potentially applicable to other dynamic forecasting/longer-term horizon decision contexts, including macro-economics (in central banking systems), energy systems and energy markets broadly, investment planning and decisions (at company, local or regional or national government levels, as well as individuals), environmental and ecological forecasting and policy advisory contexts, and many others. Other applied context beyond the core of macroeconomic policy include environmental planning and decisions, corporate and financial decision making, and basically any form of institutional planning in which formal models for data integration to advise forecasting that decisions are ultimately reliant upon. Engineering, especially in its financial aspects, is another important aspect, from dynamic reliability assessment and preventive maintenance to project management. In the arena of global scale economics and finance, there is increasing attention to integrated economic/financial/IT/social systems and the networking aspects of risk propagation through complex interacting sub-systems links to applications.
Questions: email email@example.com