Hybrid multi-objective optimization methods for in silico biochemical system production

• Main Author: Mohd Arfian, Ismail
• Format: Thesis
• Language: English
• Published: 2016
• Subjects: RA Public aspects of medicine
• Online Access: http://umpir.ump.edu.my/id/eprint/17668/16/Hybrid%20multi-objective%20optimization%20methods%20for%20in%20silico%20biochemical%20system%20production.pdf; http://umpir.ump.edu.my/id/eprint/17668/

In silica approach for multi-objective constraint optimization of biochemical system production is a computational process that aims to improve the biochemical system production. Besides the biochemical system production, the component concentrations involved also need to be considered. The optimization process involves the process of altering and ne-tuning of components in the biochemical system. The optimization process becomes complicated and difcult when a large biochemical system with many components is involved. In addition, the optimization process involves multi-objective problem which maximizes the biochemical system production and at the same time minimizing the total of component concentrations involved. Beside that, several constraints of biochemical system which are the steady state condition and component concentration constraint also contribute to the complication and difculty in optimization process. This study aims to design and develop an optimization method that efciently and effectively maximizes the biochemical system production and minimizes the total component concentrations involved simultaneously. To achieve this goal, an improved method was proposed known as Advance Newton Strength Pareto Cooperative Genetic Algorithm. The proposed method combined Newton method, Strength Pareto approach, Cooperative Coevolutionary Algorithm (CooCA) and Genetic Algorithm (GA). The use of Newton method is for dealing with biochemical system, Strength Pareto approach is for the multi-objective problem, GA is to maximize the production, and CooCA is to minimize the total component concentrations involved. The effectiveness of the proposed method was evaluated using two benchmark case studies. The experimental results showed that the proposed method was able to generate the highest results compared to other studies. Statistical validation conrmed that the proposed method is competent in producing good results in terms of maximizing the biochemical system production and minimizing the total of component concentrations involved. In conclusion, this study has presented an improved optimization method, capable to simultaneously maximize the biochemical system production and minimize the total of component concentrations involved.