The combined effect of optimal control and swarm intelligence on optimization of cancer chemotherapy. (June 2020)

• Record Type: Journal Article
• Authors: Shindi, Omar; Kanesan, Jeevan; Kendall, Graham; Ramanathan, Anand
• Is Part Of: Computer methods and programs in biomedicine. Volume 189(2020)
• Publication Date: 2020-06
• Subjects: Hybrid optimal control -- Multi objective optimization -- Constrained multi-objective optimization -- Cancer chemotherapy -- Particle swarm optimization -- Evolutionary algorithms; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2020.105327
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: Optimal control with swarm and evolutionary hybrid methodologies demonstrate that they are superior to pure swarm intelligence or evolutionary algorithm methodologies in terms of minimization of tumor and drug results and consumes far less computational time. Indirect method of optimal control with Augmented Lagrangian aids in constraints handling relieving swarm and evolutionary algorithms to focus on tumor and drug minimization. Optimal control with swarm and evolutionary hybrid methodologies results show close proximity to chemotherapy protocol maximum tolerated dose (MTD). Abstract: Background and Objectives: In cancer therapy optimization, an optimal amount of drug is determined to not only reduce the tumor size but also to maintain the level of chemo toxicity in the patient's body. The increase in the number of objectives and constraints further burdens the optimization problem. The objective of the present work is to solve a Constrained Multi- Objective Optimization Problem (CMOOP) of the Cancer-Chemotherapy. This optimization results in optimal drug schedule through the minimization of the tumor size and the drug concentration by ensuring the patient's health level during dosing within an acceptable level. Methods: This paper presents two hybrid methodologies that combines optimal control theory with multi-objective swarm and evolutionary algorithms and compares the performance of these methodologies with multi-objective swarm intelligence algorithms such as MOEAD, MODE, MOPSO and M-MOPSO. The hybrid and conventional methodologies are compared by addressing CMOOP. Results: The minimized tumor and drug concentration results obtained by the hybrid methodologies demonstrate that they are not only superior to pure swarm intelligence or evolutionary algorithm methodologies but also consumes far less computational time. Further, Second Order Sufficient Condition (SSC) is also used to verify and validate the optimality condition of the constrained multi-objective problem. Conclusion: The proposed methodologies reduce chemo-medicine administration while maintaining effective tumor killing. This will be helpful for oncologist to discover and find the optimum dose schedule of the chemotherapy that reduces the tumor cells while maintaining the patients' health at a safe level.