Applications Of Hybrid Invasive Weed Optimization Algorithm In Parameter Optimization

With the development of modern technology,many fields will use intelligent algorithm to solve practical optimization problems.Therefore,many researchers are also committed to the research and development of intelligent optimization algorithm.Optimization method plays an important role in solving engineering problems.In the past decades,researchers have used some methods inspired by biological and natural systems to solve complex optimization problems.Among the existing intelligent optimization algorithms,Most of the intelligent optimization algorithms are proposed by scholars through observing the evolution,aggregation and other behaviors of creatures in nature Optimization is a new numerical random optimization algorithm.Its idea comes from the phenomenon of weed growth,reproduction,competition and elimination in the development of agricultural production.Weeds are strong and resistant,and have strong reproductive capacity,so they show the behavior of invasion and growth.Therefore,in 2006 Mehrabian and Lucas proposed IWO optimization algorithm.As a heuristic algorithm,IWO algorithm has the advantages of strong global search ability,few parameters,easy operation,easy implementation and strong robustness.IWO optimization algorithm is widely used in many academic fields.IWO algorithm has been used to solve a variety of complex optimization problems.In this paper,four hybrid IWO algorithms are used to solve parameter inversion problem,pharmacokinetic parameter optimization problem,Muskingum model parameter optimization problem and soil water characteristic curve parameter optimization problem.In order to solve the problem of parameter inversion of solar shadow model with integer variables,an improved invasive weed optimization algorithm(HIWO)based on Quasi Newton algorithm(bfsg)is proposed.Our algorithm can not only maximize the local search ability of BFGS algorithm and the global search ability of invasive weed optimization algorithm,but also further improve the optimization ability of the algorithm by using the improved strategy with learning from the optimal individual.In addition to the parameter inversion of the sun shadow model,we also use 12 benchmark functions to verify the calculation accuracy and convergence speed of HIWO algorithm.In the benchmark function experiment,HIWO algorithm can not only achieve higher calculation accuracy than other comparison algorithms,but also show a strong competitiveness in the convergence speed,that is,the convergence speed is faster.In the experiment of parameter inversion of sun shadow model,HIWO algorithm can not only successfully reverse the date of sun shadow model,but also overcome the disadvantage of classical mathematical method that it is difficult to solve the integer nonlinear optimization problem by integer of some random variables in the algorithm.Experimental results show that HIWO algorithm not only has high accuracy,but also has fast convergence speed.HIWO can effectively improve the accuracy and efficiency of the sun shadow location technology,as well as an effective and efficient technology to deal with integer parameter inversion in engineering applications.In view of the limitation of the sensitivity of the initial value and the inability of the evolutionary algorithm to determine the search range of the traditional methods for estimating the pharmacokinetic parameters,this paper proposes a hybrid invasive weed optimization(HJIWO)algorithm which combines Hooke Jeeves(HJ)and adaptive invasive weed optimization(AIWO).Finally,we verify the performance of the HJIWO algorithm proposed in this paper through two experiments: benchmark function and two compartment model.In the benchmark function experiment,we select 15 benchmark functions.Through the analysis of experimental data,we find that HJIWO algorithm is more competitive and superior than other algorithms in terms of calculation accuracy and convergence speed.In the two compartment model experiment of extravascular administration,HJIWO algorithm is used to optimize the parameters.We can see that HJIWO is not only superior to the traditional FM algorithm in numerical stability,but also superior to HJ and IWO in error minimization.The experimental results show that HJIWO algorithm is a feasible method to solve the pharmacokinetic parameters,which has higher accuracy and stronger robustness compared with other technologies.In order to solve the problem of parameter optimization of Muskingum model,IWO has slow convergence speed and poor calculation accuracy.In this paper,a hybrid invasive weed optimization algorithm(PIWO)based on Powell algorithm and global oriented optimization strategy is proposed.This hybrid algorithm uses the local search ability of Powell algorithm to initialize the population,so that the seed can get a better solution during initialization.In the evolutionary process,the basic evolutionary strategies with global oriented optimization strategy and invasive weed optimization algorithm will play a corresponding role,which can further improve the calculation accuracy and convergence speed of the algorithm.In order to make the optimization performance of PIWO algorithm more convincing,this paper not only uses the 16 benchmark function,but also the parameter optimization problem of maskingen model.In the benchmark function experiment,PIWO algorithm can get more accurate results and faster evolution process.In the simulation experiment of the parameter optimization of the muskingen model,the experimental results show that the PIWO algorithm has high convergence accuracy and relatively fast convergence speed.This provides a new and effective method for solving the parameter optimization problem of maskingen linear model.In the field of soil water movement,soil water characteristic curve is a very important parameter.So far,van Genuchten equation(VG)is the most widely used equation of soil water characteristic curve.The problem of VG equation parameter calculation is transformed into a nonlinear optimization problem.Then,based on the basic IWO algorithm,Levy flight is introduced to estimate the parameters.Through the analysis of the simulation results,it can be concluded that the IWO hybrid algorithm based on Levy flight is better than the hybrid genetic algorithm and stochastic particle swarm optimization algorithm in solving the VG equation parameter estimation problem.In addition,we also use 15 benchmark functions to further verify the performance superiority of LIWO algorithm.In this part of experiments,we find that the calculation accuracy of LIWO algorithm is far higher than other comparison algorithms by analyzing the experimental data,and we can also find that LIWO algorithm also shows obvious competitiveness when comparing the evolution process curve,that is,convergence speed Faster.These experimental data show that the LIWO algorithm has higher accuracy and stronger robustness.