CN-121977567-A - Multi-constraint optimization method of ant colony algorithm in transformer substation path planning

CN121977567ACN 121977567 ACN121977567 ACN 121977567ACN-121977567-A

Abstract

The invention discloses a multi-constraint optimization method of an ant colony algorithm in transformer substation path planning, which comprises the steps of optimizing an pheromone updating mechanism, introducing elite ant strategy, establishing a multi-constraint adaptive model, carrying out path fine adjustment by combining a genetic algorithm, adopting a dynamic pheromone updating mechanism, adjusting the pheromone volatilization rate according to a searching progress, accelerating the exploration speed at the initial stage of the algorithm, enhancing the reinforcement of a high-quality path at the later stage of the searching, improving the convergence precision, introducing elite ant strategy, guiding the ant colony to search towards the global optimal direction through the reinforcement update of a historical optimal solution, further improving the quality of path planning, constructing the multi-constraint adaptive model, and obtaining a qualified path through the normalization processing and the punishment mechanism of constraint conditions in the process of ant colony optimization by combining environment constraint, terrain adaptability and shortest distance optimization factors. Compared with the pheromone updating mechanism of the traditional ant colony algorithm, the method can find a better path faster and reduce the calculation time.

Inventors

HUANG YAOQUN
WANG CHANGJUN
FANG YONGFENG
WANG HAILEI
ZHANG QINGQING

Assignees

安徽优航遥感信息技术有限公司

Dates

Publication Date: 20260505
Application Date: 20260126

Claims (7)

1. A multi-constraint optimization method of an ant colony algorithm in transformer substation path planning is characterized by comprising the steps of optimizing an pheromone updating mechanism, introducing elite ant strategy, establishing a multi-constraint adaptive model, and carrying out path fine adjustment by combining a genetic algorithm; Wherein: A dynamic pheromone updating mechanism is adopted, the volatile rate of the pheromone is adjusted according to the searching progress, and the searching speed is increased in the initial stage of an algorithm; Strengthening the high-quality path in the later search period, and improving convergence accuracy; Introducing elite ant strategy, and guiding ant colony to search in the global optimal direction through strengthening and updating the history optimal solution so as to further improve the quality of path planning; Constructing a multi-constraint adaptive model, and acquiring a qualified route by combining environmental constraint, terrain adaptability and shortest distance optimization factors in the ant colony optimization process through normalization processing of constraint conditions and a punishment mechanism; Fine tuning is carried out on the local optimal path by combining a genetic algorithm, so that the finally planned path is better; And carrying out additional strengthening on the historical optimal path, wherein an updating formula is as follows: ; Wherein, the For the optimal path length of the current iteration, Is the strengthening coefficient of elite ants, Represents the pheromone concentration on the path (i, j), Is the total amount of pheromones.
2. The multi-constraint optimization method of ant colony algorithm in transformer substation path planning according to claim 1, wherein in the path searching stage, ant colony starts from an initial node, and selects a forward direction according to pheromone concentration and path heuristic information by simulating ant foraging behavior, wherein each ant is located at a current position Selecting a forward node Probability of (2) Calculated from the following formula: ; Wherein, the Represents the pheromone concentration on the path (i, j), Heuristic information representing paths is typically taken I.e., the reciprocal of the path length, ensures that ants prefer shorter paths, Is a set of alternative paths for the current location of the ant, And The influence weights of the pheromone and heuristic information in path selection are respectively determined.
3. The multi-constraint optimization method of ant colony algorithm in transformer substation path planning according to claim 2, wherein the dynamic pheromone updating mechanism comprises local updating and global updating; the local updating is carried out when each ant moves to a new node, the purpose is to properly adjust the concentration of pheromone in the searching process, and the exploratory property of the path is maintained, and the updating formula is as follows: ; Wherein, the Is the volatilization rate of the pheromone, Initially setting a pheromone concentration; The global updating is performed after each iteration is finished, and is mainly used for strengthening pheromone of a high-quality path so as to accelerate convergence speed, and the updating formula is as follows: ; Wherein, the From ants Incremental determination of contributed pheromones: ; Wherein, the In the case of the total amount of pheromones, Is ant The path length travelled.
4. A multi-constraint optimization method of ant colony algorithm in transformer substation path planning according to claim 3, wherein elite ant strategy is to additionally strengthen the historical optimal path when globally updating pheromone.
5. The multi-constraint optimization method of ant colony algorithm in transformer substation path planning according to claim 4, wherein the multi-constraint adaptive model evaluates path quality by weighting multi-objective functions, including path length constraint, task priority constraint and unmanned plane self constraint, and the fitness calculation formula is as follows: ; Wherein, the For the path length of the optical fiber, As a constraint factor of the priority of the task, The factors are constrained for the drone itself, Is a weight parameter.
6. The multi-constraint optimization method of the ant colony algorithm in the transformer substation path planning according to claim 5, wherein in the final stage of path optimization, a genetic algorithm is combined to finely adjust a local optimal path, and the genetic algorithm further optimizes the high-quality path optimized by the ant colony algorithm by simulating natural selection, crossing and mutation processes so as to enable the high-quality path to meet the inspection requirement; Firstly, adopting a path node sequence coding mode, taking an optimal path obtained by an ant colony algorithm as an initial population, and calculating an fitness function: ; Wherein, the For the path length of the optical fiber, As a constraint factor of the priority of the task, The factors are constrained for the drone itself, Is a weight parameter; the higher the fitness is, the higher the probability of being selected into the next generation is, and the selection operation adopts a roulette method, so that individuals with higher fitness are more likely to inherit, and certain randomness is reserved; The interleaving operation employs partial map interleaving (PMX), exchanging partial paragraphs of paths to generate new paths; The mutation operation optimizes the feasibility of the path through two inspection points in the random exchange path, and improves the searching diversity of the algorithm; after multiple iterations, the genetic algorithm continuously optimizes the path.
7. The multi-constraint optimization method of the ant colony algorithm in the path planning of the transformer substation according to claim 6, wherein a gridding map of a substation inspection area is constructed in an initialization stage, path nodes and adjacent relations are defined, the map comprises conventional road information, topography, environment protection areas and regulation limiting constraint conditions, parameters of ant colony scale m, pheromone volatilization rate ρ and pheromone updating parameters of alpha and beta are set, the pheromone volatilization rate ρ determines the residual pheromone concentration on the path, and the parameters of alpha and beta control the relative weight of the pheromone and the path distance.

Description

Multi-constraint optimization method of ant colony algorithm in transformer substation path planning Technical Field The invention relates to the technical field of unmanned aerial vehicle control systems, in particular to a multi-constraint optimization method of an ant colony algorithm in transformer substation path planning. Background In power system planning and construction, substation inspection path planning is an important link for ensuring high efficiency, safety and economy of power transmission. Reasonable path planning can reduce engineering cost, reduce influence on environment and improve the running reliability of the power grid. Currently, path planning problems are often modeled as combinatorial optimization problems involving a variety of constraints, such as line length, terrain complexity, construction costs, environmental protection requirements, regulatory constraints, and the like. The existing path planning method mainly comprises a shortest path algorithm (such as Dijkstra algorithm and A algorithm) based on graph theory and a meta heuristic algorithm (such as genetic algorithm, particle swarm optimization algorithm and ant colony algorithm) based on intelligent optimization. The traditional shortest path algorithm is simple in calculation and suitable for the shortest path search of a single target, but cannot effectively solve the problem of multi-constraint optimization, and complex factors such as topography, regulations and environment are easily ignored, so that a planning result does not have practical feasibility. The intelligent optimization algorithm can adapt to complex multi-objective optimization problems, solves the global optimal solution in a large-scale search space, but has higher calculation complexity, is easy to sink into local optimal, has slower convergence speed, and has insufficient adaptability under the multi-constraint condition. The ant colony algorithm (AntColonyOptimization, ACO) has been widely used in the field of path planning due to its strong self-organizing optimizing and global searching capabilities. However, the application of the standard ant colony algorithm in substation inspection path planning still has a certain problem. Firstly, under the conditions of larger search space and more constraint conditions, the standard ant colony algorithm usually needs longer time to find a proper path, and has slower convergence speed. Second, due to the pheromone update mechanism, the algorithm may concentrate on certain local paths early in the search, resulting in a final solution that deviates from global optimum. Furthermore, the existing ant colony algorithm has limited multi-factor optimization processing capacity for topography, environmental protection, regulation constraint and the like, and cannot effectively balance various objective functions. In addition, in the substation inspection path optimization process, the shortest path or the lowest cost is simply targeted, so that a line can pass through an ecological protection area, a geological disaster area or a high land use area, and feasibility and compliance of engineering implementation are affected. In the prior art, CN119457629a, this patent proposes a welding robot welding path optimization method based on the ant colony algorithm, which solves the path planning problem by using the basic ant colony optimization algorithm. However, the pheromone updating mechanism of the method is relatively fixed, and the algorithm is easy to fall into local optimum due to the fact that the self-adaptive adjustment is not carried out. In the prior art, CN103279674A is a ship search and rescue method based on an ant colony algorithm, when the algorithm falls into local optimum, an update pheromone strategy ② is adopted to update the pheromone concentration on a local optimum solution path only. The ship search and rescue method solves the problems that the existing ship is difficult to accurately reach a search and rescue place due to the influence of faults, the path proficiency of operators, marine environment parameters and the like. Aiming at the problems, the invention provides a multi-constraint optimization method for improving an ant colony algorithm. According to the method, through a dynamic pheromone updating mechanism, the volatile rate of the pheromone is adjusted according to the searching progress, the convergence speed is improved, and the local optimal trap is reduced. And an elite ant strategy is introduced, and the searching capability of an optimal path is improved through global optimal reinforcement of elite ants. Meanwhile, a multi-constraint adaptation mechanism is adopted, and factors such as environmental constraint, terrain adaptability and shortest distance constraint are combined, so that a multi-constraint optimization model is established, and a path planning result is more reasonable. In addition, the invention also combines a genetic algorithm to finely adjust the loca