Novel swarm intelligent optimization algorithm-pigeon swarm algorithm

Abstract

The invention belongs to the field of optimization algorithms and provides a novel swarm intelligent optimization algorithm-pigeon swarm algorithm. The algorithm comprises a takeoff process, a flight process and a homing process. The takeoff process comprises an initialization sub-process, a soaring sub-process and a rising sub-process, which are used for initializing a pigeon swarm position, a flight speed and an optimal solution direction; the flight process comprises a flat flight sub-process, a turning sub-process and a chasing sub-process, which are used for searching for a local optimal solution and a global optimal solution and improving a global worst solution; and the homing process prevents the algorithm from falling into the local optimal solution. The algorithm has the following characteristics: 1) the algorithm has low requirement on the property of a target function, and a function expression or an expression form that is not a function form can be accepted; 2) the algorithm has the characteristics of relatively high global convergence, low algorithm cycle frequency and high convergence speed for a low-dimensional function; and 3) the algorithm has relatively high global convergence, relatively low cycle frequency and relatively high stability for a high-dimensional, multi-peak-value and complex problem.

Description

technical field [0001] The invention belongs to the field of optimization algorithms, which is used for solving the global numerical optimal solution of continuous functions, and is a novel swarm intelligence optimization algorithm: pigeon swarm algorithm. Background technique [0002] The extremum problem of a function is one of the important problems in mathematics. At present, optimization algorithms are mainly divided into two categories. One is traditional optimization algorithms, such as Newton's method, simplex method, conjugate gradient method, interval algorithm, pattern search method, branch and bound method, and filling function method. The other is swarm intelligence optimization algorithms based on the development of biology, physics and artificial intelligence, such as genetic algorithm, particle swarm algorithm, ant colony algorithm, harmony algorithm, fish swarm algorithm and monkey swarm algorithm. When traditional optimization algorithms deal with nonlinea...

AI Technical Summary

Problems solved by technology

[0003] At present, swarm intelligence optimization algorithms have played a very important role in various fields, but swarm intelligence optimization algorithms generally have premature convergence, weak global optimization ability, many algorithm cycles, and slow convergence speed in high-dimensional situations. It is easy to fall into problems such as local optima

Images