Biochemical and environmental systems are non-linear systems, sometimes its
dynamical model may not be well understood, and often data available are sparse and
subject to large error and can also be limited by seasonal constraints. Often, exemplar
data is available in conjunction with qualitative knowledge about the system's dynamics.
Which typically involves many states, complex dynamics, and the exemplar data is sparse
and noisy. This project will focus on understanding how such dynamical information can
be used to constrain the system parameter space of feasible solutions as well as
determining regions with invariant dynamics subject to parameter and structure
perturbations. The theory will be applied to the problems of model structure selection,
parameter estimation and robust experiment design. Two types of modeling philosophies
will be investigated, one with parametric dynamical model, and one based on model free
approach for systems with less knowledge on its systems dynamics. A number of cases
studies involving biochemical and environmental data will be used to validate this work.
This project is expected to have a significant impact in terms of developed theory,
improved computational tools and from demonstrating the benefit using two topical
examples of complex systems modeling. The use of qualitative knowledge about the
states' dynamics is important in systems biology because of the extensive biological
knowledge and hypotheses about the oscillatory states and positive feedback
amplification. Also, the results will be an important tool to support the ecological
research and assessment at Hong Kong’s Marine Biological Park at Hoi Ha Wan.