Robust design modelling and optimisation using response surface methodology and inverse problem

Abstract

This paper focuses on solving the robust design (RD) problem that occurs in pharmaceutical studies in which output responses are measured over time. In order to handle such a situation, firstly, a customised experimental format is proposed for pharmaceutical experimental design. Using both response surface methodology (RSM) and a new inverse problem (IP) approach, a customised method of estimation also is developed to handle dynamic time-series as well as multiple responses for pharmaceutical experiments. Next, a new inverse problem-based robust design (IPRD) model is proposed based on the mean squared error (MSE) concept. For verification and comparison, a case study of a pharmaceutical experiment is conducted.