A Symbiotic Agent-Based Network Platform Linking Expert Knowledge and Machine Learning for Systemic Risk Mitigation
The World Modelers program aims to develop tools that rapidly assemble models for food security from different, unstructured resources into
qualitative models and combine these with quantitative models. The resulting models are designed to support decision making at various
levels of abstraction, enabling the analyst or decision maker to gauge direct and side effects of decisions virtually.
Three main tasks are to be fulfilled by this project:
1.Further development of the multi-Twist model for agricultural world markets, and further development of a global system for crop yield
forecasting based on LPJmL and other PIK models. Both models shall be integrated to enable short-term forecasts of gridded yields as
well as world market price fluctuations and regional storage variation for the main staple crops under different intervention and management scenarios.
2.Ensure portability of PIK models or their regularly updated outputs to facilitate model/result integration into modeling pipelines of other World
Modelers teams
3.Support of other performer teams in subject-specific questions related to PIK-based models.
