Enhanced Fruit Fly Optimization Algorithm And Its Application

Fruit fly optimization algorithm is compiled based on the foraging behavior of Drosophila.As a classic swarm intelligent optimization algorithm,it has the characteristics of simple operation,fast operation speed,and high optimization accuracy.Compared to traditional optimization methods,the fruit fly optimization algorithm is more intelligent and convenient,with certain advantages,but there are also some shortcomings.Most swarm intelligence algorithms are prone to falling into local optimizations and unstable optimization results,especially when faced with high-dimensional problems,the optimization effect of the algorithm is often poor.The fruit fly optimization algorithm studied in this paper also has similar problems.In view of the above shortcomings,this paper improves the Fruit fly optimization algorithm and applies it to some real scenes,which not only improves its learning performance but also broadens its application field.The research contents of this paper mainly include:(1)A multi-mechanism improved fruit fly optimization algorithm is proposed.First,the population is initialized with an exponential function based on the natural constant e to improve the diversity of initial solutions and the stability of the algorithm.Secondly,the disturbance coefficient is added to dynamically adjust the search step of the algorithm,enhance the ability of the algorithm to jump out of the local optimization,and improve the global optimization ability of the algorithm.Finally,the log logarithm mechanism is used for iterative optimization to improve the convergence accuracy of the algorithm.Then,the improved algorithm is applied to function optimization and engineering case optimization.(2)An adaptive fruit fly optimization algorithm is proposed.Firstly,Cubic chaotic mapping is introduced to adjust the search step size of the fruit fly population during the iterative process,improve the global search ability of the algorithm,and enhance the optimization accuracy and stability of the algorithm.Secondly,a discriminant factor is added to achieve adaptive judgment of algorithm convergence,reducing unnecessary loss of algorithm time complexity.Finally,a variable step size mechanism is introduced to improve the random step size mechanism of the fruit fly optimization algorithm,to solve the blindness and disorder of the random step size,increase the diversity of solutions in the iterative calculation process of the algorithm,and further strengthen the optimization ability of the algorithm.Subsequently,the improved algorithm was applied to the experiment of optimizing neural network to predict gasoline octane number.(3)A variable step-size fruit fly optimization algorithm is proposed.Aiming at the problem that the original algorithm uses random search step-size for iterative optimization,resulting in low efficiency and low precision of optimization,the convergence step-size mechanism is introduced to dynamically adjust the search step-size of the algorithm,balance the overall development and local optimization capabilities of the algorithm,and improve the overall optimization potential of the algorithm.Then,the improved algorithm is applied to the image enhancement experiment of optimizing nonlinear Beta transform.