الگوریتم ژنتیکی دی ان ای اصلاح شده برای برآورد پارامتر کلروفنل اکسیداسیون 2 در آب فوق بحرانی

Based on the mechanism of biological DNA genetic information and evolution, a modified DNA genetic algorithm (MDNA-GA) is proposed to estimate the kinetic parameters of the 2-Chlorophenol oxidation in supercritical water. In this approach, DNA encoding method, choose crossover operator and frame-shift mutation operator inspired by the biological DNA are developed for improving the global searching ability. Besides, an adaptive mutation probability which can be adjusted automatically according to the diversity of population is also adopted. A local search method is used to explore the search space to accelerate the convergence towards global optimum. The performance of MDNA-GA in typical benchmark functions and kinetic parameter estimation is studied and compared with RNA-GA. The experimental results demonstrate that the proposed algorithm can overcome premature convergence and yield the global optimum with high efficiency.

Supercritical water oxidation (SCWO) has been widely studied and applied to the industrial organic wastewater treatment [1], [2] and [3]. The design, optimization and advanced control of the SCWO process require rigorous kinetic model with precise parameters. Parameter estimation in kinetics is actually a sophisticated numerical optimization problem and has attracted considerable interests. In the past decades, a lot of researchers have focused on using conventional optimization approaches such as the Levenberg–Marquardt (L–M), the Gauss–Newton and the Nelder–Mead algorithms to solve optimization problems [4], [5] and [6]. But these deterministic optimization algorithms often get trapped into so called local optima in the search process and cannot yield satisfied results. Genetic algorithm (GA), based on the heredity and process of natural biological evolution, is one of the most important global optimization techniques. It has been successfully applied to many practical engineering problems [7], [8], [9], [10], [11], [12], [13] and [14]. Although GA has powerful global search ability and better performances compared with traditional optimization algorithms when handling complex optimization problems, it does exhibit some shortages such as premature convergence, poor exploitation capability and the convergence speed to the global optimum decreased considerably in the later period of evolution. Various attempts have been made to overcome these shortages. Many scholars find that the performances of GA critically rely on the representation of the solutions and the definition of the genetic operators. Therefore, new encoding methods and genetic operations have been employed in order to enhance the performances of GA. In [15], an improved genetic algorithm based on a novel selection strategy was presented to handle nonlinear programming problems. Valarmathi et al. [16] proposed a real-coded genetic algorithm for system identification and controller tuning in a nonlinear pH process. Inspired by the mechanism of the biological DNA, a RNA genetic algorithm (RNA-GA) based on DNA computing was proposed by Tao and Wang for estimating the parameters of chemical engineering processes [17]. It is claimed that the accuracy and the diversity of the population have been improved significantly by encoding chromosomes with nucleotide bases and adopting some RNA molecular operations. Wang and Wang also proposed a novel RNA genetic algorithm (NRNA-GA) and a protein inspired RNA genetic algorithm to estimate the parameters of dynamic systems [18] and [19]. Chen and Wang developed a DNA based genetic algorithm, which adopts new genetic operators, to determine 25 kinetic parameters of hydrogenation reaction successfully [20]. In order to improve the search efficiency and prevent premature convergence, we present a modified DNA genetic algorithm (MDNA-GA). Further, this method is applied to five benchmark test functions and the kinetic parameter estimation of the 2-Chlorophenol oxidation in supercritical water.

In this paper, a modified DNA genetic algorithm is proposed and tested on a wide set of typical benchmark functions with the aim to evaluate its effectiveness in the solution of optimization problems. Also, the proposed algorithm is applied to estimate the kinetic parameters of the 2-Chlorophenol oxidation in supercritical water. The computational results have shown that MDNA-GA can improve the accuracy and the convergence rate considerably than RNA-GA. It is a promising technique for solving realistic parameter estimation problems. In addition, future research efforts should focus on applying MDNA-GA to more complicated nonlinear parameter estimation problems in chemical or biological kinetics.