Solving Inverse Problems for Mechanistic Systems Biology Models with Unknown Inputs

Abstract

The goal of the proposed project is to develop and test the feasibility of a novel approach to solve the inverse problem for a class of systems arising from systems biology study, in which input is unknown (e.g. cannot be observed) but multiple outputs can be observed. As mentioned above, the most significant challenge associated with this class of inverse problems is that the standard approach to inverse problem (where the parameters of the mechanistic model are optimized to fit the input-output data) is not applicable in the absence of system input observation. To resolve this challenge, this project proposes to exploit the commonality shared by the outputs that originate from the same input. Indeed, noting that these outputs originate from the identical input, a relationship between these outputs can be formulated, which can subsequently be utilized in solving the inverse problem without necessitating the input observations.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 16, 2014
Accession Number
ADA622227

Entities

People

  • Jin-Oh Hahn

Organizations

  • University of Maryland

Tags

DTIC Thesaurus Topics

  • Biology
  • Computational Biology
  • Engineering
  • Engineers
  • Health Services
  • Inverse Problems
  • Mathematical Analysis
  • Mathematical Models
  • Measurement
  • Military Research
  • Personalized Medicine
  • Range Finding
  • Simulations
  • Standards
  • Steady State
  • Students
  • Systems Biology

Fields of Study

  • Computer science

Readers

  • Approximation Theory.
  • Computational Modeling and Simulation
  • Systems Analysis and Design