EFFICIENT IDENTIFICATION ALGORITHM FOR CONTROLLING MULTIVARIABLE TUMOR MODELS: GRADIENT-BASED AND TWO-STAGE METHOD

This paper presents Gradient-based Iterative (GI) and Two-Stage Gradient-based Iterative (2S-GI) identification algorithms for the Controlled Auto-Regressive Moving Average (CARMA) form of a multivariable tumor model. The mathematical proof of the 2S-GI algorithm for multivariable CARMA systems is provided, demonstrating its effectiveness in parameter estimation. The step-by-step introduction of the algorithm facilitates further studies and implementation. A comprehensive comparison between the GI and 2S-GI algorithms is conducted, evaluating their performance in terms of convergence rate and estimation accuracy. The introduced multivariable tumor model serves as a testbed for the algorithms’ effectiveness. The results of the comparison, supported by simulated data, demonstrate the superiority of the 2S-GI algorithm in accurately estimating the parameters of the CARMA system. This research provides valuable insights into the application of gradient-based iterative algorithms in controlling multivariable tumor models, paving the way for improved control strategies in cancer treatment.