CFE-CMStatistics 2024: Start Registration
View Submission - CFECMStatistics2024
A1212
Title: Optimal design construction: A comparative study of various methods Authors:  Akram Mahmoudi - Jönköping International Business School (Sweden) [presenting]
Saumen Mandal - University of Manitoba (Canada)
Abstract: Owing to the widespread application of optimal designs, the motivation is to compare different methods of optimum designs. In this regard, some optimization algorithms are used to construct approximate optimal designs. Some of these methods are gradient-based, while others are gradient-free algorithms. The studied methods include a class of multiplicative algorithms, simulated annealing and optimization with inequality constraints-Barrier methods. Optimization algorithms are simulated in three models: a quadratic model, a practical model in chemistry, and a model with two design variables. The simulation study is done under different scenarios to explore the performance of the algorithms. Then, the strengths and weaknesses of the methods are explored, and a comparison between methods is done.