Path planning method based on multi-strategy improved parrot optimization algorithm

Abstract

The invention discloses a path planning method based on a multi-strategy improved parrot optimization algorithm. The method comprises the following steps: initializing parameters; initializing a parrot population by adopting a mapping strategy based on a good point set, and generating a plurality of paths; calculating the fitness value of each path based on the target function, and determining the optimal path during initialization; on the basis of a random walk strategy, simulating the behaviors of the parrot to update the population, and adjusting the position of each parrot in the population according to the behaviors of the parrot; updating the optimal path and the fitness value, and judging whether a crisscross strategy is adopted or not based on a crisscross factor; if the crisscross strategy is adopted, continuing to update the optimal path and the fitness value, judging whether a termination condition is met or not, and if the termination condition is met, outputting the optimal path and the fitness value; the convergence speed and optimization precision of the algorithm are effectively improved by introducing three strategies, and the path generated by the algorithm is better in smoothness and faster in convergence speed.

Description

Technical Field

[0001] The present invention relates to a path planning method, and particularly to a path planning method based on a multi-strategy improved parrot optimization algorithm. Background Art

[0002] As a heuristic optimization method, the parrot optimization algorithm is inspired by numerous natural behaviors, especially simulating the behavioral patterns of the Molina's Red-fronted Parakeet, including foraging, staying, communicating, and fear of strangers. These behaviors are translated into mathematical models to guide the position update during the search process, aiming to find the optimal solution, thereby maximizing task efficiency, reducing resource consumption, and improving the reliability of task completion.

[0003] The parrot optimization algorithm has excellent optimization capabilities and fast convergence, and is applicable to various path planning problems. However, it often encounters challenges when exploring the global solution space, especially being prone to falling into local optimal solutions when facing complex problems. For example, the random docking point selection in the staying behavior, the probabilistic decision-making in the communication behavior, and the movement path simulation in the foraging behavior using the Lévy flight distribution. It can be applied to various path planning problems.

[0004] Although the parrot optimization algorithm has shown good performance in various applications, it still has some deficiencies. Especially in terms of global optimization ability, the algorithm is easily trapped by local optimal solutions and difficult to find the optimal solution. Summary of the Invention

[0005] Object of the Invention: The object of the present invention is to provide a path planning method to improve the convergence speed and optimization accuracy of the algorithm during robot path planning.

[0006] Technical Solution: The path planning method described in the present invention includes the following steps:

[0007] S1. Initialize parameters, set the starting position, target position, population size, maximum number of iterations, map boundary conditions, and obstacle positions of the robot;

[0008] S2. Initialize the parrot population using a mapping strategy based on the good point set to generate multiple paths;

[0009] S3. Calculate the fitness value of each path based on the objective function to determine the best path at initialization;

[0010] S4. Based on the random walk strategy, simulate the parrot behavior to update the population, and adjust the position of each parrot in the population according to the parrot's behavior, where the parrot's behavior includes foraging behavior, staying behavior, communicating behavior, and fear of strangers behavior;

[0011] S5. Update the optimal path and fitness value, and determine whether to adopt the cross strategy based on the cross factor;

[0012] S6. If the cross strategy is adopted, continue to update the optimal path and fitness value, and determine whether the termination condition is met. If the termination condition is met, output the optimal path and fitness value; otherwise, return to S4 for continued iteration.

[0013] Preferably, the mapping strategy of the good point set in S2 is as follows:

[0014]

[0015] Among them, P n (k) represents the good point set, r represents the good point, represents taking the fractional part, n represents the number of points, and x i (j) represents the optimized initial parrot population, and ub j represents the upper bound of the jth dimension, and lb j represents the lower bound of the jth dimension.

[0016] Preferably, the objective function in S3 is as follows:

[0017]

[0018] Among them, (x k , y k ) represents the coordinates of the kth node, d represents the path length of the robot from the starting point to the ending point, and n represents the number of intermediate nodes on the path.

[0019] Preferably, the random walk strategy in S4 is as follows:

[0020] X(t) = {0, Cumsum[2r(1) - 1], Cumsum[2r(2) - 1], …, Cumsum[2r(n) - 1]};

[0021]

[0022] Among them, X(t) represents the set of random walks, Cumsum represents the accumulation of random step lengths, n represents the maximum number of random walks; r(t) represents a random function, and t represents the current number of steps of random walk.

[0023] Preferably, the random walk strategy is as follows:

[0024]

[0025] Among them, and respectively represent the upper and lower boundaries of the dimth dimension in the tth iteration of the random walk, represents the position of the parrot in the dim-th dimension at the t-th iteration, S f represents the scaling factor of the random walk.

[0026] Preferably, the scaling factor of the random walk is as follows:

[0027]

[0028] where N represents the population size.

[0029] Preferably, it further includes normalizing the random walk, and the formula is as follows:

[0030]

[0031] where a i and b i represent the minimum and maximum values of the random walk of the i-th parrot in the dim-th dimension, represents the position of the i-th parrot in the dim-th dimension at the t-th iteration.

[0032] Preferably, it is characterized in that the cross-over strategy of S6 is as follows:

[0033]

[0034] where, represents the offspring obtained by cross-over, a1 and a2 represent random numbers between 0 and 1, dim represents the dimension, k represents the (i + 1)-th population, b1 represents a random number between -1 and 1, d1 and d2 represent integers of the random dimension of the population; t represents the current iteration number, and T represents the total number of iterations;

[0035] The judgment condition is: if the random number rand is greater than the probability factor P, then the cross-over strategy is adopted to perturb the optimal position of the parrot; otherwise, this strategy is skipped.

[0036] Advantageous effects: Compared with the prior art, the present invention has the following remarkable advantages: 1. By introducing the good point set mapping strategy, different from the method of setting the initial position of the parrot by using a random function in the classical parrot optimization algorithm, the population is evenly distributed at the initial stage; 2. By introducing the random walk strategy, intervention is carried out according to the rules during the population iteration stage to prevent it from falling into the local optimal solution and increase the global optimization ability of the classical parrot optimization algorithm; 3. By introducing the cross-over strategy, cross-learning between different individuals in the same dimension is promoted, the search blind area is reduced, and the global search ability is enhanced; 4. By introducing the three strategies, the convergence speed and optimization accuracy of the algorithm can be effectively improved, and the smoothness of the path generated by the algorithm is better and the convergence speed is faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Schematic diagram of the process of the present invention;

[0038] Figure 2 Schematic diagram of the path planning simulation result in a simple environment of the embodiment of the present invention;

[0039] Figure 3 Schematic diagram of the comparison of the path planning convergence curves in a simple environment of the embodiment of the present invention;

[0040] Figure 4 Schematic diagram of the path planning simulation result in a complex environment of the embodiment of the present invention;

[0041] Figure 5 Schematic diagram of the comparison of the path planning convergence curves in a complex environment of the embodiment of the present invention. Detailed implementation manners

[0042] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings.

[0043] S1. Initialize parameters, set the starting position S and the target position G of the robot, the population size N and the maximum number of iterations T, and determine the map boundary conditions and the positions of obstacles.

[0044] S2. Initialize the parrot population by using the mapping strategy based on the good point set and generate multiple paths.

[0045] The principle of the good point set adopted is:

[0046] Let G s be the unit cube in the s-dimensional Euclidean space. If r ∈ G s , in the form of:

[0047]

[0048] where the deviation satisfies where C(r,ε)n -1+ε is a constant related only to r and ε (ε is an arbitrary positive number), then P n (k) is called a good point set and r is called a good point.

[0049] represents taking the fractional part, n represents the number of points, taking p is the smallest prime number that satisfies .

[0050] Mapping it to the search space is:

[0051]

[0052] The above formula is the optimized formula for initializing the parrot population, ubj represents the upper bound of the j-th dimension, lb j represents the lower bound of the j-th dimension.

[0053] S3. Calculate the fitness value of each path based on the objective function to determine the optimal path at initialization.

[0054] The objective function is:

[0055]

[0056] where (x k , y k ) represents the coordinates of the k-th node, d represents the path length of the robot from the starting point to the ending point, and n represents the number of intermediate nodes on the path.

[0057] S4. Based on the random walk strategy, simulate the behavior of parrots to update the population, and adjust the position of each parrot in the population according to the behavior of the parrots. The behaviors of the parrots include foraging behavior, staying behavior, communication behavior, and fear behavior towards strangers.

[0058] The communication behavior and fear behavior towards strangers of the parrot algorithm adopt the random walk strategy as follows:

[0059] X(t) = {0, Cumsum[2r(1) - 1], Cumsum[2r(2) - 1], …, Cumsum[2r(n) - 1]};

[0060] where X(t) represents the set of random walks, Cumsum represents the accumulation of random step sizes, n represents the maximum number of random walks, and r(t) represents a random function, where t represents the current number of steps of the random walk;

[0061] r(t) is:

[0062]

[0063] The position of the parrot is determined by the good point set, but the position of each parrot will affect the boundary of its random walk.

[0064] The upper and lower boundaries of the random walk are as and shown:

[0065]

[0066]

[0067] and represent the upper and lower boundaries of the dim-th dimension in the t-th iteration of the random walk respectively, Represents the position of the parrot in the dim-th dimension at the t-th iteration;

[0068] S f Represents the scaling factor of the random walk, and the formula is as follows:

[0069]

[0070] Because the search of the PO algorithm is restricted by upper and lower bounds, the random walk needs to be normalized, and the formula is as follows:

[0071]

[0072] a i and b i Represent the minimum and maximum values of the random walk of the i-th parrot in the dim-th dimension; Represents the position of the i-th parrot in the dim-th dimension at the t-th iteration.

[0073] S5. Update the best path and fitness value, and determine whether to adopt the cross strategy based on the cross factor.

[0074] The cross probability factor P is:

[0075]

[0076] t represents the current iteration number, and T represents the total number of iterations.

[0077] If the random number rand is greater than the probability factor P, the cross strategy is adopted to perturb the optimal position of the parrot; otherwise, this strategy is skipped.

[0078] S6. If the cross strategy is adopted, continue to update the best path and fitness value, and determine whether the termination condition is satisfied. If the termination condition is satisfied, output the best path and fitness value; otherwise, return to S4 to continue the iteration.

[0079] The foraging behavior and staying behavior of the parrot algorithm adopt the cross strategy, and the formula is as follows:

[0080]

[0081] Among them, Represents the offspring through cross, a1 and a2 represent random numbers between 0 and 1, dim represents the dimension, k represents mod(i, N)+1, that is, the current population i takes the remainder of the population N, namely the i+1-th population; b1 represents a random number between -1 and 1, and d1 and d2 represent integers of the random dimension of the population.

Claims

1. A path planning method based on a multi-strategy improved Parrot optimization algorithm, characterized in that: The following steps are involved: S1. Initialize parameters, set the robot's starting position, target position, population size, maximum number of iterations, map boundary conditions, and obstacle positions; S2, using a mapping strategy based on a good point set to initialize the parrot population and generate multiple paths; S3. Calculate the fitness value of each path based on the objective function and determine the optimal path at the time of initialization; S4. Based on the random walk strategy, the parrot behavior is simulated to update the population, and the position of each parrot in the population is adjusted according to the parrot behavior, wherein the parrot behavior includes foraging behavior, staying behavior, communication behavior, and fear of strangers; S5, updating the optimal path and fitness value, and determining whether to adopt the vertical and horizontal crossover strategy based on the vertical and horizontal crossover factor; S6. If the vertical and horizontal cross strategy is adopted, the optimal path and fitness value are continuously updated to determine whether the termination condition is met. If the termination condition is met, the optimal path and fitness value are output. Otherwise, the process returns to S4 to continue iteration.

2. The path planning method according to claim 1, wherein The mapping strategy of the good point set described in S2 is as follows: Among them, P n (k) represents the good point set, r represents the good point, Indicates taking the decimal part, n indicates the number of points, x i (j) represents the optimized initialized parrot population, ub j represents the upper bound of the j-dimensional space, lb j represents the lower bound of the j-dimension.

3. The path planning method according to claim 1, wherein: The objective function of S3 is as follows: Among them, (x k , y k ) represents the coordinates of the k-th node, d represents the path length of the robot from the starting point to the ending point, and n represents the number of intermediate nodes on the path.

4. The path planning method according to claim 1, characterized in that The random walk strategy described in S4 is as follows: X(t)={0,Cumsum[2r(1)-1],Cumsum[2r(2)-1],…,Cumsum[22r(n)-1]}; Among them, X(t) represents the set of random walks, Cumsum represents the accumulation of random steps, and n represents the maximum number of random walks; r(t) represents a random function, and t represents the number of steps in the current random walk.

5. The path planning method according to claim 4, characterized in that: The random walk strategy is as follows: Among them, and respectively represent the upper and lower boundaries of the random walk in the dim-th dimension at the t-th iteration, represents the position of the parrot in the dim-th dimension at the t-th iteration, S f represents the scaling factor of the random walk.

6. The path planning method according to claim 5, wherein The scaling factors of the random walk are as follows: Where N represents the population size.

7. The path planning method according to claim 4, wherein It also includes normalization of random walks, the formula is as follows: where a i and b i represent the minimum and maximum values of the i-th parrot's random walk in the dim-th dimension, represents the position of the i-th parrot in the dim-th dimension at the t-th iteration.

8. The path planning method according to claim 1, characterized in that: The crossover strategy described in S6 is as follows: in, represents the offspring through vertical and horizontal crossover, a1 and a2 represent random numbers between 0 and 1, dim represents the dimension, k represents the i+1th population, b1 represents a random number between -1 and 1, d1 and d2 represent integers of the random dimension of the population; t represents the current number of iterations, and T represents the total number of iterations; The judgment condition is: if the random number rand is greater than the probability factor P, the vertical and horizontal cross strategy is used to disturb the optimal position of the parrot; otherwise, the strategy is skipped.

9. A computer device, characterized in that: The method comprises one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and when the programs are executed by the processors, the steps of the path planning method according to claims 1-8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the steps of the path planning method according to claims 1 to 8 are implemented.

Images