Optimization With Julia


Creative Commons License

Satman M. H., Akadal E.

in: Global Studies on Management Information Systems, Elif Kartal,Emre Akadal,Gökhan Övenç,Saeed Tabar, Editor, Istanbul University Press, İstanbul, pp.295-315, 2023

  • Publication Type: Book Chapter / Chapter Research Book
  • Publication Date: 2023
  • Publisher: Istanbul University Press
  • City: İstanbul
  • Page Numbers: pp.295-315
  • Editors: Elif Kartal,Emre Akadal,Gökhan Övenç,Saeed Tabar, Editor
  • Istanbul University Affiliated: Yes

Abstract

Management science is concerned mainly with optimization techniques, although it is not so apparent to everyone. Determining the number of products in stocks, establishing an appropriate number of payment points, assigning the most suitable job to the right worker, selecting the best production order in multiple machines, and developing an optimal transportation strategy, are all classical optimization problems in management. Optimization is a particular sub-discipline of mathematics that serves as an underlying infrastructure for Operations Research, Statistics, and Machine Learning. For example, a substantial range of tools, from the least-squares estimator and mathematical programming to deep learning and feed-forward neural networks are based on minimizing an objective function with a great number of optimization parameters. These optimization problems are linear or nonlinear. The problem at hand may have constraints as well. When the decision variables are in types of integer or binary, things get even worse in the computational context. Julia is a high-performance programming language primarily designed for scientific computing. Even though Julia is a relatively young language, it has several packages for optimization, data analysis, and machine learning. In addition, with its dynamic and robust type system, Julia is promising in terms of performance as a compiled language. In this chapter, we investigate many optimization problems, e.g., linear and nonlinear objective functions, with or without constraints, integer or mixed types of decision variables, etc., and search for the ability to solve them using Julia. Benchmark statistics are also reported to present the performance of Julia in solving the mentioned problems.