Computational Optimization Methods in Statistics, Econometrics and Finance

WP 1: Optimization and numerical methods

• Development of a general data driven optimization heuristic and comparison with other multipurpose methods (1,6)
• Development of hybrid heuristics for estimation of parameters of (nonlinear) time series models and comparison with standard techniques (3,12)
• Development and analysis of optimization heuristics for optimal control and dynamic game solutions in stochastic models (6,11)
• Heuristic optimization algorithms for high-breakdown point regression estimates (6)
• Defining standards for the presentation of results of (stochastic) optimization algorithms; analysis of required information (1,6,10)
• High performance implementations (parallel, distributed and grid computing), e.g., for indirect estimation of model parameters and general simulation (6,9,10) 

WP 2: Statistical analysis of stochastic optimization tools

• Statistical analysis of the convergence of optimization heuristics for uniform design problems: theoretical description and empirical findings (regression analysis) (1,10)
• Evaluation of optimization heuristics by means of extreme value theory (1, 6) – Analysis of joint convergence of heuristics and econometric estimates (1,8,10)
• Statistical inference in evolutionary forecasting methods (3)
• Impact of automatic model selection tools on the statistical analysis of time series models, e.g., vector error correction models and non-linear time series models (1,9,12) 

WP 3: Identification of promising fields of applications, implementation and evaluation of new (heuristic) optimization tools

• Genetic algorithms and other heuristic methods for non-linear model building in time series analysis, e.g., regime switching and conditional heteroskedastic models (3,12)
• Using optimization heuristics to derive joint bootstrap intervals for impulse response analysis of vector error correction models (1,12)
• Optimal model selection in linear and nonlinear time series models (8,9,10)
• Heuristic algorithms for dynamic optimization (optimum control, dynamic games) and their application to macroeconometric models (11,12)
• Optimization based decision support system for macroeconomic policy analysis (11,12)
• Risk management and dynamic trading strategies, financial engineering and hedging strategies, applications in insurance and reinsurance (7,10)
• Estimation of latent factor models for high-frequency data (7,8)
• Calibration methodologies for derivative pricing models: Stochastic and local volatility models (with jumps); pricing measure; hybrid derivatives (2,4,7)
• Minimization of explicit time dependence in option pricing models (2,4,5) 

WP 4: Modelling of artificial markets

• Analysis of reinforcement learning and genetic algorithms for applications of agent based models of market design and for the analysis of general economic modelling (10,11)
• Implementation and application of a standard method for the evaluation of agent based models based on an indirect estimation approach (1,6,10)
• Improving the modelling of auction and market design in agent based models based on evaluation results (7,8,10)
• Establishing an evaluation standard for agent based models (1,10,11)