CN105760954A - Parking system path planning method based on improved ant colony algorithm - Google Patents

Abstract

The invention discloses a parking system path planning method based on an improved ant colony algorithm, and aims at solving the problem of AGV vehicle access path planning in an intelligent parking garage so that vehicle accessing can be completed in the shortest possible time, utilization rate of parking places can be enhanced, time of waiting for vehicle accessing can be reduced for social members and automatic management of parking equipment can be realized. The concrete planning steps are that an AGV working environment model in the intelligent parking garage is created by adopting a grid method; the conventional ant colony algorithm is optimized and improved by introducing of new node state transfer probability and an updating strategy of combination of local and global pheromones; and simulated testing is performed on the AGV vehicle access path planning process by applying the improved ant colony algorithm and the result is outputted. The method has high global search capability and great convergence performance, and can effectively enhance path search efficiency, shorten search path length and reduce the number of path turnings and can also enable the AGV to effectively avoid obstacles in the complex operation environment so as to search the optimal collision-free path.

Description

A kind of based on the parking system paths planning method improving ant group algorithm

Technical field

The present invention relates to a kind of based on the parking system paths planning method improving ant group algorithm, belong to AGV Path Planning Technique field.

Background technology

Sharply increasing of automobile pollution highlights with parking position demand surplus contradiction day by day along with parking position is under-supply, causes the social problem such as urban traffic congestion, parking difficulty, becomes the bottleneck of restriction urban economy and social development.And based on the appearance of AGV horizontal mobile intelligent parking garage, then solve this difficult problem well.Compared with conventional planar garage and mechanical stereo garage, this intelligent garage has that parking floor space bicycle parking quantity few, effective is many, vehicle access automaticity high, cost performance is high and security reliability advantages of higher, it may be achieved the various functions such as unmanned automatically storing and taking vehicles, AGV automatic charging and garage automatic charging.The core in studying plane movable-type intelligent garage parking is to solve AGV Transport Vehicle path planning problem.At present, for path planning problem, relevant scholar has done a large amount of research work both at home and abroad, and in succession proposes tabu search algorithm, simulated annealing, genetic algorithm and particle cluster algorithm etc..These algorithms have respective advantage, but there is also many defects, as complicated in algorithm, be easily absorbed in that local optimum, search volume be big and search efficiency etc..

Ant group algorithm is a kind of modern Bio-simulated Evolution algorithm proposed by the Food Recruiment In Ants behavior of observation of nature circle the nineties in 20th century by Italy scholar Dorigo.M et al., this algorithm has stronger robustness, it is easily achieved parallel processing, and easily with the advantage such as other heuritic approaches are combined, it is widely used in solving the optimization problem of different field, such as traveling salesman problem, routing issue, Job-Shop problem and vehicle and robot path planning's problem etc., and achieve good effect.But this algorithm there is also program runtime length, convergence rate and is absorbed in the defects such as local optimum slowly and easily.For solving intelligent parking garage AGV Transport Vehicle path planning problem, strengthen algorithm ability of searching optimum, improve route searching efficiency, accelerate algorithm the convergence speed, shortening searching route length, the present invention proposes a kind of based on the parking system paths planning method improving ant group algorithm.

A kind of difference based on the parking system paths planning method with a kind of method existence essence improving ant colony algorithm optimization support vector machine parameter proposed such as Zhang Li that improve ant group algorithm that the present invention proposes.Both identical points are all on the basis of Basic Ant Group of Algorithm, by the pluses and minuses of analysis of classical ant group algorithm, propose respectively to adopt different optimisation strategy that classical ant group algorithm has been improved, and are used for improvement ant group algorithm solving to produce practical problem.Difference is in that, the improvement ant group algorithm that Zhang Li etc. propose is by finding out its maximum and minima in object function solution, and on this basis by adopting different global information element more new formulas respectively the worst solution of globally optimal solution and the overall situation to be strengthened and weakened.Although adopting the method can improve the speed and accuracy rate of seeking best of breed, but can there is certain blindness in its search procedure, searching route there will be deviation, convergence rate and is absorbed in the defects such as locally optimal solution slowly and easily.And the present invention has considered traditional ant group algorithm search blindness, constringency performance difference and has easily been absorbed in the defects such as local optimum, realize the Optimal improvements to tradition ant group algorithm by introducing new node state transition probability and pheromone local and the overall more New Policy combined.The present invention is adopted to contribute to the strengthening Formica fusca purposiveness to route searching, it is ensured that Formica fusca is more guiding in path search process, also can strengthen the ability of searching optimum of algorithm simultaneously, improve convergence of algorithm performance etc..

Summary of the invention

For solving in intelligent stereo garage AGV Transport Vehicle path planning problem and overcoming tradition ant group algorithm convergence rate slowly, to be easily absorbed in the defects such as locally optimal solution, the invention provides a kind of based on the parking system paths planning method improving ant group algorithm.The present invention, compared with tradition ant group algorithm, can be effectively improved route searching efficiency, reduces algorithmic statement algebraically, shortens searching route length, reduces path turnover number of times and increase substantially searching route quality.

Parking system paths planning method based on improvement ant group algorithm provided by the invention is mainly made up of three parts: adopt Grid Method to create the working environment model of AGV in intelligent parking garage；By introducing new node state transition probability and pheromone local and the overall more New Policy combined, tradition ant group algorithm is optimized improvement；Use improvement ant group algorithm that AGV Transport Vehicle path planning process is carried out emulation testing and exports result.

For achieving the above object, the technical solution adopted in the present invention is based on the parking system paths planning method improving ant group algorithm, specifically includes following steps:

Step 1, employing Grid Method create the running environment model of AGV in intelligent garage, and particular content is as follows:

1-1): AGV and running environment are handled as follows: (a) AGV running environment must be reduced to two-dimensional finite space；B the Obstacle Position in () Projected Operational Environment is it is known that can use For Polygons Representation, and ignore its short transverse；C () assumes that AGV travels at the uniform speed in two-dimensional finite space, ignore it and turn to, lift and the factor such as parked in short-term；D () AGV is reduced to a particle；E () AGV can only take the air line, it is impossible to walk the diagonal of grid.

1-2): utilizing photographic head, radar sensor, infrared ray sensor and laser sensor etc. that intelligent garage and AGV carry to gather AGV running environment information, above-mentioned information includes the initial parking stall of AGV, target parking stall, barrier and AGV position to be charged etc..

1-3): using the environmental information of aforesaid operations collection as modeling data, Grid Method is adopted to create AGV two dimension running environment model；

Step 2, the initialization each parameter of ant group algorithm

Initialized parameter is needed to specifically include that the beginning and end position of AGV, Formica fusca quantity m, the initial value NC and maximum NC_max of iterations, pheromone track significance level factor of influence α, node transfer expected degree factor-beta, pheromone volatility coefficient ρ and constant Q.

Step 3, node state transition probability being improved by introducing new distance heuristic function, and path pheromone is carried out local updating, particular content is as follows:

3-1): for improving the Formica fusca purposiveness to route searching, improve the ability of searching optimum of algorithm and search efficiency, the present invention using the inverse of Euclidean distance between next node j and destination node g as heuristic function.

The distance heuristic function of tradition ant group algorithm is as follows:

$d(i,j)=\sqrt{{(x_i-x_j)}^2+{(y_i-y_j)}^2}$

η_ij=1/d (i, j)

In formula, (x_i,y_j) and (x_j,y_j) represent i, j two coordinate figure of node respectively；(i j) represents node (i, the Euclidean distance between j) to d；η_ijRepresent the expected value between present node i and next node j.

The distance heuristic function improving ant group algorithm is as follows:

$d(j,g)=\sqrt{{(x_j-x_g)}^2+{(y_j-y_g)}^2}$

η_jg=1/d (j, g)

In formula, (x_j,y_j) and (x_g,y_g) represent j, g two coordinate figure of node respectively；(j g) represents node (j, the Euclidean distance between g) to d；η_jgRepresent the expected value between next node j and destination node g.

3-2): the improvement distance heuristic function factor is incorporated into node state transition probability computing formula

The node state transition probability of tradition ant group algorithm is:

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[{\mathrm{\η}}_{ij}\left(t\right)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[{\mathrm{\η}}_{ij}\left(t\right)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

The node state transition probability improving ant group algorithm is:

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(j,g)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{is}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(s,g)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

In formula,Represent that Formica fusca k is at the internodal transition probability of i, j；allowed_kRepresent the optional node set of Formica fusca k；τ_ijRepresent that Formica fusca k is retained in path (i, the pheromone on j).

3-3): if certain moment Formica fusca k is positioned at node i, then the transition probability between itself and next node j can be obtained by following formula calculating:

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(j,g)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{is}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(s,g)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

3-4): calculated by above formula and obtain the present node i set with next node j transition probability, in conjunction with roulette method, it is determined that next node j.

The limit that 3-5): after next node j determines, Formica fusca just need to have been passed by (i, j) carries out pheromone local updating, and local message element more new formula is as follows:

τ_ij(t+1)=(1-ρ) τ_ij(t)+ρτ₀

In formula, τ₀For the pheromone under initial condition, ρ represents pheromone volatility coefficient, ρ ∈ [0,1].

Step 4, Pheromone update mode being improved by introducing pheromone local and the more New Policy that combines of the overall situation, and path pheromone carries out overall situation renewal, particular content is as follows:

4-1): for guaranteeing that Formica fusca has more directiveness in path search process, strengthen the ability of searching optimum of algorithm and improve convergence of algorithm performance, by introducing new pheromone incremental computations formula, global information element more new formula being improved.

Tradition ant group algorithm global information element more new formula is as follows:

τ_ij(t+1)=(1-ρ) τ_ij(t)+Δτ_ij

${\mathrm{\Δ\τ}}_{ij}=\underset{k=1}{\overset{m}{\Σ}}{\mathrm{\Δ\τ}}_{ij}^k$

In formula, Δ τ_ijRepresent that Formica fusca k is retained in path (i, the pheromone increment on j)；ρ represents pheromone volatility coefficient (ρ ∈ [0,1])；L_kRepresent the path that Formica fusca k current iteration searches；Q is constant, represents pheromone concentration enhancer.

Improve ant group algorithm global information element more new formula as follows:

τ_ij=(1-ρ) τ_ij+ρΔτ_ij

In formula, τ_ij(i, j) pheromone on limit, ρ represents pheromone volatility coefficient, and span is [0,1] to be retained in path for Formica fusca；Δτ_ijIt is retained in path this (i, j) the pheromone increment on limit for Formica fusca；L_gbRepresent the shortest path length that order Ant Search up to now arrives.

4-2): judge whether all Formica fuscas complete current iteration route searching, if so, then go to step 4-3), otherwise go to the 3-3 in step 3)；

4-3): add up whole path optimizings that all Formica fuscas can search up to now, choose that wherein length is the shortest one, calculate Formica fusca in the overall situation updates be retained in this path (i, j) the pheromone increment on limit, computing formula is as follows:

4-4): the routing information element increment obtained is substituted into following formula, and is realized the overall situation renewal of this shortest path pheromone by following formula.

τ_ij=(1-ρ) τ_ij+ρΔτ_ij

Step 5, judge whether iterations meets requirement, if meeting requirement, then exporting result, otherwise, then going to step 3.

Beneficial effect

(1) adopt Grid Method to create AGV running environment model, there is precision height, be easily achieved and the advantages such as accurate solution can be obtained；

(2) by introducing new distance heuristic function, node state transition probability is improved, contribute to the strengthening Formica fusca purposiveness to route searching, improve ability of searching optimum and the search efficiency of algorithm；

(3) adopt the more New Policy that pheromone local and the overall situation combine that Pheromone update mode is improved, it is advantageously ensured that Formica fusca has more directiveness in path search process, avoid Formica fusca to converge to same path, and then strengthen and improve ability of searching optimum and the constringency performance of algorithm；

(4) it is applied to based in the intelligent garage parking system path planning of AGV by improving ant group algorithm, AGV effective avoiding obstacles in complicated running environment can be made then to search a nothing and to touch optimal path, and this algorithm can effectively shorten searching route length at path search process, accelerate convergence rate, improve route searching efficiency, reduce path turnover number of times, improve searching route quality.

Accompanying drawing explanation

Fig. 1 is algorithm flow chart；

Fig. 2 is the AGV running environment model adopting Grid Method to create at random；

Fig. 3 is the running orbit iteration diagram of AGV under 10 × 10 grid environment；

Fig. 4 is the running orbit iteration diagram of AGV under 15 × 15 grid environment；

Fig. 5 is the running orbit iteration diagram of AGV under 20 × 20 grid environment.

Detailed description of the invention

For showing objects and advantages of the present invention further, below in conjunction with accompanying drawing and example, the present invention is further elaborated, but the present invention is not limited to this example.

The present invention is to provide a kind of based on the parking system paths planning method improving ant group algorithm.Accompanying drawing 1 show inventive algorithm implementing procedure figure, the flow chart describes the solution procedure of optimal path, and particular content comprises the steps:

Step 1: adopting Grid Method to create the running environment model of AGV in intelligent garage, particular content is as follows:

1-1): AGV and running environment are handled as follows: (1) AGV running environment must be reduced to two-dimensional finite space；(2) Obstacle Position in Projected Operational Environment is it is known that can use For Polygons Representation, and ignores its short transverse；(3) assume that AGV travels at the uniform speed in two-dimensional finite space, ignore it and turn to, lift and the factor such as parked in short-term；(4) AGV is reduced to a particle；(5) AGV can only take the air line, it is impossible to walks the diagonal of grid.

1-2): utilizing photographic head, radar sensor, infrared ray sensor and laser sensor etc. that intelligent garage and AGV carry to gather AGV running environment information, above-mentioned information includes the initial parking stall of AGV, target parking stall, barrier and AGV position to be charged etc..

1-3): using the environmental information of aforesaid operations collection as modeling data, Grid Method is adopted to create AGV two dimension running environment model；

Accompanying drawing 2 show the AGV running environment model adopting Grid Method to create at random, and in figure, white grid represents free grid, and AGV can free-running operation in this region；Black grid represents obstacle grid, and AGV may not operate in this region；The starting point of S and T respectively AGV and aiming spot.

Step 2: parameter initialization.

Parameter is provided that Formica fusca quantity m=40；α=5；β=9；ρ=0.1；Q=200；Maximum iteration time NC_max=80；Starting point coordinate S respectively (0.5,9.5), (0.5,14.5) and (0.5,19.5)；Terminal point coordinate E respectively (9.5,0.5), (14.5,0.5) and (19.5,0.5).

Step 3: by introducing new distance heuristic function, node state transition probability being improved, and path pheromone is carried out local updating, particular content is as follows:

3-1): design new distance heuristic function and be introduced in node state transition probability computing formula:

$d(j,g)=\sqrt{{(x_j-x_g)}^2+{(y_j-y_g)}^2}$

η_jg=1/d (j, g)

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(j,g)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{is}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(s,g)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

In formula, (x_j,y_j) and (x_g,y_g) represent j, g two coordinate figure of node respectively；(j g) represents node (j, the Euclidean distance between g) to d；Represent that Formica fusca k is at the internodal transition probability of i, j；allowed_kRepresent the optional node set of Formica fusca k；τ_ijRepresent that Formica fusca k is retained in path (i, the pheromone on j)；η_jgRepresent the expected value between next node j and destination node g.

3-2): if certain moment Formica fusca k is positioned at node i, then the transition probability between itself and next node j can be obtained by following formula calculating:

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(j,g)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{is}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(s,g)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

3-3): calculated by above formula and obtain the present node i set with next node j transition probability, in conjunction with roulette method, it is determined that next node j.

The limit that 3-4): after next node j determines, Formica fusca just need to have been passed by (i, j) carries out pheromone local updating, and local message element more new formula is as follows:

τ_ij(t+1)=(1-ρ) τ_ij(t)+ρτ₀

In formula, τ₀For the pheromone under initial condition, ρ represents pheromone volatility coefficient, ρ ∈ [0,1].

Step 4: crossing and introduce pheromone local and Pheromone update mode is improved by the more New Policy that combines of the overall situation, and path pheromone carries out overall situation renewal, particular content is as follows:

4-1): design new pheromone increment formula, and be introduced in the routing information element overall situation more new formula:

τ_ij=(1-ρ) τ_ij+ρΔτ_ij

In formula, τ_ij(i, j) pheromone on limit, ρ represents pheromone volatility coefficient, and span is [0,1] to be retained in path for Formica fusca；Δτ_ijIt is retained in path this (i, j) the pheromone increment on limit for Formica fusca；L_gbRepresent the shortest path length that order Ant Search up to now arrives.

4-2): judge whether all Formica fuscas complete current iteration and then reach home, if then going to 4-3), the 3-2 in step 3 is otherwise gone to)；

4-3): add up whole path optimizings that all Formica fuscas can search up to now, choose that wherein length is the shortest one, calculate Formica fusca in the overall situation updates be retained in this path (i, j) the pheromone increment on limit, computing formula is as follows:

4-4): the routing information element increment obtained is updated to following formula, and is realized the overall situation renewal of this shortest path pheromone by following formula.

τ_ij=(1-ρ) τ_ij+ρΔτ_ij

Step 5: judge whether iterations meets NC >=80, if meeting, then exports optimum results, otherwise, then goes to the 3-2 in step 3).

Feasibility and the effectiveness of ant colony AGV Transport Vehicle path planning in intelligent stereo garage is improved for checking, under the grid environment of 10 × 10,15 × 15 and 20 × 20, selecting tradition ant group algorithm and improvement ant group algorithm that AGV Transport Vehicle path planning process has been carried out emulation testing respectively, test result is such as shown in Fig. 3, Fig. 4 and Fig. 5.

Under the grid environment of 10 × 10, simulation result is as shown in Figure 3.In figure, solid black lines represents the Transport Vehicle running orbit improving AGV under ant group algorithm and path iteration change curve respectively, and black dotted lines represents running orbit and the path iteration change curve of AGV under tradition ant group algorithm respectively.In path, tradition ant group algorithm and improvement ant group algorithm are 18m；In number of times is transferred in path, tradition ant group algorithm is 9 times, and improving ant group algorithm is 6 times；In algorithmic statement algebraically, tradition ant group algorithm starts convergence in the 13rd generation, and improvement ant group algorithm is 2nd generation.

Under the grid environment of 15 × 15, simulation result is as shown in Figure 4.In figure, solid black lines represents the Transport Vehicle running orbit improving AGV under ant group algorithm and path iteration change curve respectively, and black dotted lines represents running orbit and the path iteration change curve of AGV under tradition ant group algorithm respectively.In path, tradition ant group algorithm is 30m, and improvement ant group algorithm is 28m；In number of times is transferred in path, tradition ant group algorithm is 14 times, and improving ant group algorithm is 8 times；In algorithmic statement algebraically, tradition ant group algorithm starts convergence in the 37th generation, and improving ant group algorithm was the 6th generation.

Under the grid environment of 20 × 20, simulation result is as shown in Figure 5.In figure, solid black lines represents the Transport Vehicle running orbit improving AGV under ant group algorithm and path iteration change curve respectively, and black dotted lines represents running orbit and the path iteration change curve of AGV under tradition ant group algorithm respectively.In path, tradition ant group algorithm is 40m, and improvement ant group algorithm is 38m；Path turnover number of times aspect, tradition ant group algorithm is 14 times, and improving ant group algorithm is 11 times；In algorithmic statement algebraically, tradition ant group algorithm starts convergence in the 43rd generation, and improving ant group algorithm was the 13rd generation.

Following table show under the grid environment of 10 × 10,15 × 15 and 20 × 20, by selecting tradition ant group algorithm and the emulation data improved ant group algorithm that AGV Transport Vehicle path planning process is carried out emulation testing and obtain.

Under different size grid environment, by AGV Transport Vehicle path planning process is carried out Multi simulation running experiment, can as drawn a conclusion:

(1) tradition ant group algorithm and improvement ant group algorithm all can make AGV effectively avoid the barrier in surrounding, then search one and touch path optimizing from the nothing of origin-to-destination；

(2) compared with tradition ant group algorithm, ant group algorithm optimizing path is improved the shortest, path turnover least number of times, start convergence times minimum；

(3) iteration diagram display in path improves ant group algorithm and has stronger ability of searching optimum, good constringency performance, faster search speed and higher search efficiency.

It is only that principles of the invention and core concept have been elaborated and illustrated described in examples detailed above; the interest field of the present invention can not be limited with this; should be understood that; for those skilled in the art; under the premise without departing from principle of the present invention; also can make some deformation, and these deformation also should be regarded as protection scope of the present invention.

Claims (4)

1. the parking system paths planning method based on improvement ant group algorithm, it is characterised in that comprise the steps:

Step 1, employing Grid Method create the running environment model of AGV in intelligent garage；

Step 2, the initialization each parameter of ant group algorithm:

Initialized parameter is needed to include: the beginning and end position of AGV, Formica fusca quantity m, the initial value NC and maximum NC_max of iterations, pheromone track significance level factor of influence α, node transfer expected degree factor-beta, pheromone volatility coefficient ρ and constant Q；

Step 3, node state transition probability is improved by introducing new distance heuristic function, and path pheromone is carried out local updating；

Step 4, Pheromone update mode is improved by introducing pheromone local and the more New Policy that combines of the overall situation, and path pheromone is carried out overall situation renewal；

Step 5, judge whether iterations meets requirement, if meeting requirement, then exporting result, otherwise, then going to step 3.

2. the method for claim 1, it is characterised in that in step 1, comprises the following specific steps that:

1-1), AGV and running environment are handled as follows: (a) AGV running environment must be reduced to two-dimensional finite space；B the Obstacle Position in () Projected Operational Environment is it is known that can use For Polygons Representation, and ignore its short transverse；C () assumes that AGV travels at the uniform speed in two-dimensional finite space, ignore it and turn to, lift and the factor such as parked in short-term；D () AGV is reduced to a particle；E () AGV can only take the air line, it is impossible to walk the diagonal of grid；

1-2), utilizing photographic head, radar sensor, infrared ray sensor and laser sensor etc. that intelligent garage and AGV carry to gather AGV running environment information, above-mentioned information includes the initial parking stall of AGV, target parking stall, barrier and AGV position to be charged etc.；

1-3), using the environmental information of aforesaid operations collection as modeling data, employing Grid Method creates AGV two dimension running environment model.

3. the method for claim 1, it is characterised in that in step 3, particular content is as follows:

3-1), for improving the Formica fusca purposiveness to route searching, improve the ability of searching optimum of algorithm and search efficiency, using the inverse of Euclidean distance between next node j and destination node g as heuristic function；

The distance heuristic function of tradition ant group algorithm is as follows:

$d(i,j)=\sqrt{{(x_i-x_j)}^2+{(y_i-y_j)}^2}$

η_ij=1/d (i, j)

In formula, x_i,y_jAnd x_j,y_jRepresent i, j two coordinate figure of node respectively；(i j) represents node (i, the Euclidean distance between j) to d；η_ijRepresent the expected value between present node i and next node j；

The distance heuristic function improving ant group algorithm is as follows:

$d(j,g)=\sqrt{{(x_j-x_g)}^2+{(y_j-y_g)}^2}$

η_jg=1/d (j, g)

In formula, x_j,y_jAnd x_g,y_gRepresent j, g two coordinate figure of node respectively；(j g) represents node (j, the Euclidean distance between g) to d；η_jgRepresent the expected value between next node j and destination node g；

3-2): the improvement distance heuristic function factor is incorporated into node state transition probability computing formula

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(j,g)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{is}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(s,g)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

In formula,Represent that Formica fusca k is at the internodal transition probability of i, j；allowed_kRepresent the optional node set of Formica fusca k；τ_ijRepresent that Formica fusca k is retained in the pheromone on path i, j；

3-3): if certain moment Formica fusca k is positioned at node i, then the transition probability between itself and next node j can be obtained by following formula calculating:

$p_{ij}^k\left(t\right)=\left\{\begin{array}{cc}\frac{{\[{\mathrm{\τ}}_{ij}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(j,g)\]}^{\mathrm{\β}}}{\underset{s\∈{\mathrm{allowed}}_k}{\Σ}{\[{\mathrm{\τ}}_{is}\left(t\right)\]}^{\mathrm{\α}}\·{\[1/d(s,g)\]}^{\mathrm{\β}}},& j\∈{\mathrm{allowed}}_k\\ 0,& otherwise\end{array}\right.$

3-4): calculated by above formula and obtain the present node i set with next node j transition probability, in conjunction with roulette method, it is determined that next node j.；

3-5): after next node j determines, limit i, the j that need to Formica fusca have just been passed by carries out pheromone local updating, and local message element more new formula is as follows:

τ_ij(t+1)=(1-ρ) τ_ij(t)+ρτ₀

In formula, τ₀For the pheromone under initial condition, ρ represents pheromone volatility coefficient, ρ ∈ [0,1].

4. the method for claim 1, it is characterised in that in step 4, particular content is as follows:

4-1), for guaranteeing that Formica fusca has more directiveness in path search process, strengthen the ability of searching optimum of algorithm and improve convergence of algorithm performance, by introducing new pheromone incremental computations formula, global information element more new formula being improved；

τ_ij=(1-ρ) τ_ij+ρΔτ_ij

In formula, τ_ijBeing retained in the pheromone on i, j limit, path for Formica fusca, ρ represents pheromone volatility coefficient, and span is 0,1；Δτ_ijIt is retained in the pheromone increment on this i, j limit, path for Formica fusca；L_gbRepresent the shortest path length that order Ant Search up to now arrives；

4-2), judge whether all Formica fuscas complete current iteration route searching, if so, then go to step 4-3), otherwise go to the 3-3 in step 3)；

4-3), add up whole path optimizings that all Formica fuscas can search up to now, choose that wherein length is the shortest one, calculate Formica fusca in the overall situation updates and be retained in the pheromone increment on i, j limit, this path, computing formula is as follows:

4-4), the routing information element increment that obtain is substituted into following formula, and realizes the overall situation of this shortest path pheromone by following formula and update,

τ_ij=(1-ρ) τ_ij+ρΔτ_ij。