CN106681135A - Cable wiring route searching method based on mixed water drop algorithm - Google Patents

Abstract

The invention discloses a cable route automatic planning method based on an improved mixed water drop algorithm. The method mainly solves the problems that in the prior art, the algorithm efficiency is not high and a route is not smooth. The planning steps are as follows: modeling a cable wiring environment; optimizing parameters in an intelligent water drop algorithm by means of a particle swarm algorithm, and an orientation is provided for the movement of water drops; improving the water drop movement mode in the water drop algorithm, and conducting route optimization; according to the optimization result, outputting the cable route. The method considers the stability and robustness of the algorithm comprehensively, improves the search efficiency of the algorithm, avoids the occurrence of interference between cables and obstacles, and reduces the cable length required by cable wiring.

Description

A kind of cable routing pathway searching method based on mixing water droplet algorithm

Technical field

The present invention relates to self routing field, more particularly to a kind of cable routing pathway search based on mixing water droplet algorithm Method.

Background technology

The search of cable routing pathway is one of the key technology for realizing cable self routing, can rapidly find out one just Victory, collisionless, routing path that is smooth and being attached at wall not only improve the self routing efficiency of cable, while also improving The reliability of electronic product.

At present, commonly used method for searching path has Visual Graph method, HEURISTIC ALGORITHM FOR GRAPH SEARCH, Artificial Potential Field Method etc., this A little algorithms have respective pluses and minuses, and such as Artificial Potential Field Method has good real-time, but there is trap area, and close The shortcomings of can not finding path between barrier, algorithm is more suitable for solving single-object problem.Between last decade, with artificial The research of intelligent algorithm constantly makes progress, and many intelligent algorithms are also used in the path planning of mobile robot, including mould Fuzzy logic and enhancing learning algorithm, neutral net, genetic algorithm and ant group algorithm etc..These algorithms have respective advantage, But problems are there is also, such as algorithm the convergence speed is excessively slow, be easily absorbed in local optimum, search space is excessive etc..

Although being improved to many algorithms at this stage, preferable cable routing pathway can be found out, still be deposited Iterationses it is more, operation time is long the problems such as, these problems cannot meet the requirement of cable self routing, and obtain Path in, cable trails be in vacant state, this will have a strong impact on the stability of electronic product, at the same reduce electronic product work Make reliability.

The content of the invention

For the shortcoming of prior art, the present invention provides a kind of cable routing pathway search based on mixing water droplet algorithm and calculates Method, the method iterationses are few, operation time is short, can meet the requirement of cable self routing, do not interfere with the steady of electronic product It is qualitative, improve electronic product functional reliability.

Realizing the technical scheme of the object of the invention is：

A kind of cable routing pathway searching method based on mixing water droplet algorithm, comprises the steps：

1) in wiring space, the wall that cable lays is chosen, wiring space is simplified to into one in the way of planar development The routing planes of individual rectangle；

2) grating map modeling is carried out to rectangular wiring plane；

3) using particle cluster algorithm is as the coarse search stage and carries out fast path optimizing, by the multigroup sub-optimal path for obtaining Update SOIL DISTRIBUTION；

4) using improved intelligent water drop algorithm as the fine searching stage, optimum path search is carried out；

5) using the node in path and inflection point as control point, and according to step 1) in planar development mode by two The control point coordinate transformation of dimension is the Three dimensions control point coordinates in wiring space；

6) using three rank B-spline curves path optimizings；

By above-mentioned steps, using the parameter in particle cluster algorithm Intelligent Optimal water droplet algorithm, the mobile offer of the drop that feeds water It is oriented to；The move mode of water droplet in water droplet algorithm is improved, and carries out optimum path search；B-spline curves path optimizing is finally adopted, it is comprehensive Conjunction considers the stability and robustness of algorithm, can rapidly find out that one convenient, collisionless, wiring that is smooth and being attached at wall Path.Cable wiring is completed by laying cable on the path.

Step 1) in, in wiring space, the wall that cable lays is chosen, by wiring space letter in the way of planar development The routing planes of one rectangle of chemical conversion：

First, it is if the plane non-rectangle plane for simplifying after by planar development that plane extension is rectangular flat Face；

Secondly, expansion is considered as barrier；

Finally, the plane is divided into into two regions：Barrier region and non-barrier region.

Step 2) in, grating map modeling is carried out to rectangular wiring plane：

First, grating map is made up of obstacle grid with non-obstacle grid；

Secondly, arbitrarily the grid containing barrier is considered as obstacle grid；

Finally, if wiring terminal is located at the same row in grating map, grating map ranks are exchanged.

Step 3) in, it is multigroup by what is obtained using particle cluster algorithm is as the coarse search stage and carries out fast path optimizing Sub-optimal path updates SOIL DISTRIBUTION, specifically includes：

First, initial parameter (population scale N, inertia weight w, Studying factors c are set₁And c₂, maximum iteration time Ic, speed Degree maximum V_max, and speed minima V_min)；

Then, the random position vector x for generating particle_i=(x_i1,x_i2,…,x_iD) and velocity v_i=(v_i1,v_i2,…, v_iD), and initialize the locally optimal solution p of particle_pAnd globally optimal solution p_g；

Into iteration and the positional information and velocity information of all particles are updated according to following two formula：

I=1 in formula, 2 ..., N, d=1,2 ..., D, w are inertia weight, and c1 and c2 is Studying factors, r1 and r2 be [0, 1] relatively independent between, equally distributed random number, p_p=(p_p1,p_p2,…,p_pD) for the locally optimal solution of particle, i.e. particle exists Optimal solution in last iteration, p_g=(p_g1,p_g2,…,p_gD) for particle globally optimal solution, i.e., particle this iteration it Front optimal solution.

Meanwhile, calculate the fitness f (x of each particle_i), if f is (x_i) be minima in this iteration, then p_p=x_iIf, Now f (x_i)<f(p_g), then p_g=x_i, wherein fitness function f is：

(x in formula_j,y_j) for path node j coordinate, n for path point number, num_jFor node j and node j-1 it Between path touch the quantity of obstacle grid, pun causes the path for touching obstacle to give poor evaluation of estimate for penalty factor.

Finally, into next iteration recurring formula (1) (2) (3), locally optimal solution p is updated_pAnd the overall situation most solves p_g, directly Maximum is reached to iterationses or meet required precision.

Step 4) in, using improved intelligent water drop algorithm as the fine searching stage, optimum path search is carried out,

First, quantity W of water droplet, maximum iteration time Is, initial soil amount Its, to each position in grating map are set Soil initialized, and the T groups suboptimal solution obtained using PSO algorithms updates soil, following two formula：

Soil in formula_iFor the amount of soil of a certain node i；pos_iFor the position of the node；path_subFor path in D groups path In each node position；Eup is the beeline between two adjacent cells central points；λ_iExpression has λ in this T group suboptimal solution_iGroup Solution contains node i；Parameter δ, α and β are positive number.

Then, all water droplets are placed in connecting up starting point.

Path point selection, and record path nodal information are carried out using roulette method, formula is as follows：

Soil in formula_iFor the amount of soil of a certain node；pos_iFor the position of the node；Eup is two adjacent cells central points Between beeline；P is the probability for selecting next node；N is the set of the optional grid of next step；Function dis (pos_i, pos_j) for euclidean metric between node i and node j；ρ is the positive number less than 1；a_f, b_fAnd c_fIt is positive number.

When water droplet moves to another node j, the speed of water droplet is updated：

Vel in formula_jFor the speed that water droplet moves to node j；soil_jFor the amount of soil of node j；a_vFor positive number.

Meanwhile, update the soil of moved to node, and the soil that now water droplet is carried, following two formula：

Pos in formula_iFor the position of node i；Function dis (pos_i,pos_j) for Euclid between node i and node j Tolerance；soil_jFor the amount of soil of node j；ρ₁And a_sIt is positive number.

Before all water droplets move to wiring terminal, recurring formula (6) (7) (8) (9) (10) calculates each water droplet institute The fitness in path is formed, and updates locally optimal solution p_pWith globally optimal solution p_g。

After all water droplets move to wiring terminal, global amount of soil is updated：

Soil in formula_iFor the amount of soil of node i；pos_iFor the position of node i；To form the water of locally optimal solution The entrained amount of soil of drop；p_pFor locally optimal solution, and p_gFor globally optimal solution；The quantity of grid in num grating maps；Locally Perception factor ρ₂With global knowledge factor ρ₃It is positive number, and ρ₃≥ρ₂。

After complete office's amount of soil is updated according to formula (11), algorithm completes an iteration.Hereafter algorithm will be entered down An iteration, all water droplets will be refitted in wiring starting point, and method described above (formula (6) to formula (11)) re-starts shifting It is dynamic, until iterationses reach maximum or meet required precision.

Step 5) in, using the node in path and inflection point as control point, and according to step 1) in planar development The control point coordinate transformation of two dimension is the Three dimensions control point coordinates in wiring space by mode；

Step 6) in, using three rank B-spline path optimizings：

First obtaining step 5) in, control point p₀(x₀,y₀,z₀),...,p_n(x_n,y_n,z_n), determine the exponent number of B-spline curves K=3, defines the support Interval T=[t of SPL₀,t₁,...,t_n+k] such as following formula：

Then the basic function N of SPL is calculated_i,k(t), following two formula：

T is the variable in SPL equation in formula；T=[t₀,t₁,...,t_n+k] for SPL support Interval.

The equation of three rank B-spline curves, such as following formula can finally be tried to achieve：

In formula, t is the variable of curvilinear equation；N_i,kT () is the basic function in B-spline curves；

The curve exponent number that k=3 is.

Beneficial effect：

The invention provides a kind of cable routing pathway searching method based on mixing water droplet algorithm, the method：

(1) wall is laid by choosing cable, wiring space is reduced to into routing planes, after cable trails are obtained, will The two-dimensional coordinate at cable control point is converted into the three-dimensional coordinate of wiring space in plane, so that routing path is posted on wall On, it is to avoid cable hanging situation；

(2) using grating map modeling pattern, there is high precision, is easily achieved；

(3) particle cluster algorithm is got up with the fusion of improved intelligent water drop algorithm.The advantage of particle cluster algorithm is its phase It is very short the time required to for intelligent water drop algorithm, it is added in before intelligent water drop algorithm and not only will not consume too many time, shadow Overall operation speed, and the pheromone feedback that initial soil distribution can be provided for intelligent water drop algorithm are rung, intelligence can be so reduced Can water droplet algorithm run time, can reduce search time and get back the effect of path optimizing so as to reach；

(4) using euclidean metric as a factor for affecting water droplet to select node motion, so that water droplet to the greatest extent may be used Convergence of algorithm speed can be increased towards the preferable region movement in position；

(5) using three rank B-spline curves, so that cable trails are more smoothed, while avoid in cable trails depositing In the inflection point that a large amount of bent angles are larger.

Description of the drawings

Fig. 1 grating map Establishing process figures；

Fig. 2 is the cable routing pathway search routine figure based on mixing water droplet algorithm；

Fig. 3 is the cable self routing completed in wiring space by the inventive method；

Fig. 4 simplifies wiring space flow chart.

Specific embodiment

Present invention is further elaborated with reference to the accompanying drawings and examples, but is not limitation of the invention.

Embodiment

A kind of cable routing pathway searching method based on mixing water droplet algorithm, comprises the steps：

1) in wiring space, the wall that cable lays is chosen, wiring space is simplified to into one in the way of planar development The routing planes of individual rectangle；

2) grating map modeling is carried out to rectangular wiring plane；

3) using particle cluster algorithm is as the coarse search stage and carries out fast path optimizing, by the multigroup sub-optimal path for obtaining Update SOIL DISTRIBUTION；

4) using improved intelligent water drop algorithm as the fine searching stage, optimum path search is carried out；

5) using the node in path and inflection point as control point, and according to step 1) in planar development mode by two The control point coordinate transformation of dimension is the Three dimensions control point coordinates in wiring space；

6) using three rank B-spline curves path optimizings；

By above-mentioned steps, using the parameter in particle cluster algorithm Intelligent Optimal water droplet algorithm, the mobile offer of the drop that feeds water It is oriented to；The move mode of water droplet in water droplet algorithm is improved, and carries out optimum path search；B-spline curves path optimizing is finally adopted, it is comprehensive Conjunction considers the stability and robustness of algorithm, can rapidly find out that one convenient, collisionless, wiring that is smooth and being attached at wall Path.Cable wiring is completed by laying cable on the path.

Step 1) in, in wiring space, the wall that cable lays is chosen, by wiring space letter in the way of planar development The routing planes of one rectangle of chemical conversion：

As shown in figure 1, for the simplified way of wiring space.Wherein P₁To connect up starting point, P₂To connect up emphasis；S_o, S₁And S₂It is the wall that cable lays；O₁For barrier, and O₂For extended area and it is considered as barrier；Finally with the region table of black Show barrier, the non-barrier region of region representation of white.

First, it is if the plane non-rectangle plane for simplifying after by planar development that plane extension is rectangular flat Face；

Secondly, expansion is considered as barrier；

Finally, the plane is divided into into two regions：Barrier region and non-barrier region.

Step 2) in, grating map modeling is carried out to rectangular wiring plane：

As shown in Fig. 2 being grid modeling pattern.Wherein annulus represents wiring starting point, and straight-flanked ring represents wiring terminal；In it is black Colour grid lattice are obstacle grid；White grid is non-obstacle grid；The distance between any two grid for their central points company Line.

First, grating map is made up of obstacle grid with non-obstacle grid；

Secondly, arbitrarily the grid containing barrier is considered as obstacle grid；

Finally, if wiring terminal is located at the same row in grating map, grating map ranks are exchanged.

Step 3) in, it is multigroup by what is obtained using particle cluster algorithm is as the coarse search stage and carries out fast path optimizing Sub-optimal path updates SOIL DISTRIBUTION, specifically includes：

First, initial parameter (population scale N, inertia weight w, Studying factors c are set₁And c₂, maximum iteration time Ic, speed Degree maximum V_max, and speed minima V_min)；

Then, the random position vector x for generating particle_i=(x_i1,x_i2,…,x_iD) and velocity v_i=(v_i1,v_i2,…, v_iD), and initialize the locally optimal solution p of particle_pAnd globally optimal solution p_g；

Into iteration and the positional information and velocity information of all particles are updated according to following two formula：

I=1 in formula, 2 ..., N, d=1,2 ..., D, w are inertia weight, and c1 and c2 is Studying factors, r1 and r2 be [0, 1] relatively independent between, equally distributed random number, p_p=(p_p1,p_p2,…,p_pD) for the locally optimal solution of particle, i.e. particle exists Optimal solution in last iteration, p_g=(p_g1,p_g2,…,p_gD) for particle globally optimal solution, i.e., particle this iteration it Front optimal solution.

Meanwhile, calculate the fitness f (x of each particle_i), if f is (x_i) be minima in this iteration, then p_p=x_iIf, Now f (x_i)<f(p_g), then p_g=x_i, wherein fitness function f is：

(x in formula_j,y_j) for path node j coordinate, n for path point number, num_jFor node j and node j-1 it Between path touch the quantity of obstacle grid, pun causes the path for touching obstacle to give poor evaluation of estimate for penalty factor.

Finally, into next iteration recurring formula (1) (2) (3), locally optimal solution p is updated_pAnd the overall situation most solves p_g, directly Maximum is reached to iterationses or meet required precision.

Step 4) in, using improved intelligent water drop algorithm as the fine searching stage, optimum path search is carried out,

First, quantity W of water droplet, maximum iteration time Is, initial soil amount Its, to each position in grating map are set Soil initialized, and the T groups suboptimal solution obtained using PSO algorithms updates soil, following two formula：

Soil in formula_iFor the amount of soil of a certain node i；pos_iFor the position of the node；path_subFor path in D groups path In each node position；Eup is the beeline between two adjacent cells central points；λ_iExpression has λ in this T group suboptimal solution_iGroup Solution contains node i；Parameter δ, α and β are positive number.

Then, all water droplets are placed in connecting up starting point.

Path point selection, and record path nodal information are carried out using roulette method, formula is as follows：

Soil in formula_iFor the amount of soil of a certain node；pos_iFor the position of the node；Eup is two adjacent cells central points Between beeline；P is the probability for selecting next node；N is the set of the optional grid of next step；Function dis (pos_i, pos_j) for euclidean metric between node i and node j；ρ is the positive number less than 1；a_f, b_fAnd c_fIt is positive number.

When water droplet moves to another node j, the speed of water droplet is updated：

Vel in formula_jFor the speed that water droplet moves to node j；soil_jFor the amount of soil of node j；a_vFor positive number.

Meanwhile, update the soil of moved to node, and the soil that now water droplet is carried, following two formula：

Pos in formula_iFor the position of node i；Function dis (pos_i,pos_j) for Euclid between node i and node j Tolerance；soil_jFor the amount of soil of node j；ρ₁And a_sIt is positive number.

Before all water droplets move to wiring terminal, recurring formula (6) (7) (8) (9) (10) calculates each water droplet institute The fitness in path is formed, and updates locally optimal solution p_pWith globally optimal solution p_g。

After all water droplets move to wiring terminal, global amount of soil is updated：

Soil in formula_iFor the amount of soil of node i；pos_iFor the position of node i；To form the water of locally optimal solution The entrained amount of soil of drop；p_pFor locally optimal solution, and p_gFor globally optimal solution；The quantity of grid in num grating maps；Locally Perception factor ρ₂With global knowledge factor ρ₃It is positive number, and ρ₃≥ρ₂。

After complete office's amount of soil is updated according to formula (11), algorithm completes an iteration.Hereafter algorithm will be entered down An iteration, all water droplets will be refitted in wiring starting point, and re-start movement to formula (11) with above-mentioned formula (6), directly Maximum is reached to iterationses or meet required precision.

Step 5) in, using the node in path and inflection point as control point, and according to step 1) in planar development The control point coordinate transformation of two dimension is the Three dimensions control point coordinates in wiring space by mode；

Step 6) in, using three rank B-spline path optimizings：

As shown in figure 3, as using the routing path example after the optimization of three rank B-splines.Wherein wiring space is one The casing of 10dm × 18dm × 20dm, barrier is the object of dotted line, and the short cylinder that solid line is surrounded represents connector, can Will become apparent from cable connecting wiring and play not-go-end, avoiding obstacles, and apply is in wall；

First obtaining step 5) in, control point p₀(x₀,y₀,z₀),...,p_n(x_n,y_n,z_n), determine the exponent number of B-spline curves K=3, defines the support Interval T=[t of SPL₀,t₁,...,t_n+k] such as following formula：

Then the basic function N of SPL is calculated_i,k(t), following two formula：

T is the variable in SPL equation in formula；T=[t₀,t₁,...,t_n+k] for SPL support Interval.

The equation of three rank B-spline curves, such as following formula can finally be tried to achieve：

In formula, t is the variable of curvilinear equation；N_i,kT () is the basic function in B-spline curves；

The curve exponent number that k=3 is.

As shown in Figure 4：Solve routing path mixing water droplet algorithm flow be：

S101 starts

S102 initiation parameters

S103 adopts PSO Algorithm T group sub-optimal paths

S104 updates global amount of soil according to T groups sub-optimal path

S105 iterationses increase

S106 selects the water droplet for not yet moving

S107 water droplets select node motion

Soil at S108 more new nodes, water droplet translational speed, water droplet carries soil

S109 judges whether water droplet moves to wiring terminal

S110 judges whether all water droplets are moved to wiring terminal

S111 updates locally optimal solution and globally optimal solution

S112 updates global soil

S113 judges whether to meet iteration termination condition

S114 obtains control point coordinates

S115 adopts three rank B-spline curves path optimizings

S116 output cord routing paths

S117 terminates.

Claims (7)

1. it is a kind of based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that to comprise the steps：

1) in wiring space, the wall that cable lays is chosen, wiring space is simplified to into a square in the way of planar development The routing planes of shape；

2) grating map modeling is carried out to rectangular wiring plane；

3) using particle cluster algorithm is as the coarse search stage and carries out fast path optimizing, updated by the multigroup sub-optimal path for obtaining SOIL DISTRIBUTION；

4) using improved intelligent water drop algorithm as the fine searching stage, optimum path search is carried out；

5) using the node in path and inflection point as control point, and according to step 1) in planar development mode by two dimension Control point coordinate transformation is the Three dimensions control point coordinates in wiring space；

6) using three rank B-spline curves path optimizings；

By above-mentioned steps, using the parameter in particle cluster algorithm Intelligent Optimal water droplet algorithm, the mobile offer guiding of the drop that feeds water； The move mode of water droplet in water droplet algorithm is improved, and carries out optimum path search；B-spline curves path optimizing is finally adopted, is comprehensively examined The stability and robustness of algorithm are considered, can rapidly find out that one convenient, collisionless, wiring topology that is smooth and being attached at wall Footpath, by cable is laid on the path cable wiring is completed.

2. it is according to claim 1 based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that step It is rapid 1) in, in wiring space, choose the wall that lays of cable, wiring space is simplified to into a square in the way of planar development The routing planes of shape：

First, the plane is extended to into rectangle plane if the plane non-rectangle plane for simplifying after by planar development；

Secondly, expansion is considered as barrier；

Finally, the plane is divided into into two regions：Barrier region and non-barrier region.

3. it is according to claim 1 based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that step It is rapid 2) in, grating map modeling is carried out to rectangular wiring plane：

First, grating map is made up of obstacle grid with non-obstacle grid；

Secondly, arbitrarily the grid containing barrier is considered as obstacle grid；

Finally, if wiring terminal is located at the same row in grating map, grating map ranks are exchanged.

4. it is according to claim 1 based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that step It is rapid 3) in, using particle cluster algorithm is as the coarse search stage and carries out fast path optimizing, by multigroup sub-optimal path for obtaining more New SOIL DISTRIBUTION, specifically includes：

First, initial parameter (population scale N, inertia weight w, Studying factors c are set₁And c₂, maximum iteration time Ic, speed is most Big value V_max, and speed minima V_min)；

Then, the random position vector x for generating particle_i=(x_i1,x_i2,…,x_iD) and velocity v_i=(v_i1,v_i2,…,v_iD), And initialize the locally optimal solution p of particle_pAnd globally optimal solution p_g；

Into iteration and the positional information and velocity information of all particles are updated according to following two formula：

$v_{id}^{k+1}={\mathrm{wv}}_{id}^{k+1}+c_1{r}_1(p_{pd}-x_{id}^k)+c_2{r}_2(p_{gd}-x_{id}^k)---\left(1\right)$

$x_{id}^{k+1}=x_{id}^k+v_{id}^{k+1}---\left(2\right)$

Meanwhile, calculate the fitness f (x of each particle_i), if f is (x_i) be minima in this iteration, then p_p=x_iIf, now f (x_i)<f(p_g), then p_g=x_i, wherein fitness function f is：

$f=\underset{j=2}{\overset{n}{\&Sigma;}}\sqrt{{(x_j-x_{j-1})}^2+{(y_j-y_{j-1})}^2}+pun\&times;{\mathrm{num}}_j---\left(3\right)$

Finally, into next iteration recurring formula (1) (2) (3), locally optimal solution p is updated_pAnd the overall situation most solves p_g, until repeatedly Generation number reaches maximum or meets required precision.

5. it is according to claim 1 based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that step It is rapid 4) in, using improved intelligent water drop algorithm as the fine searching stage, carry out optimum path search, first, arrange water droplet quantity W, maximum iteration time Is, initial soil amount Its is initialized to the soil of each position in grating map, and is calculated using PSO The T groups suboptimal solution that method is obtained updates soil, following two formula：

$\begin{array}{ccc}{\mathrm{soil}}_i={\mathrm{soil}}_i+\frac{{\mathrm{\&lambda;}}_i}{T}\&times;\frac{\mathrm{\&delta;}\&times;Initsoil}{\mathrm{\&alpha;}+\mathrm{\&beta;}\&times;{\mathrm{eup}}^3}& if& {\mathrm{pos}}_i\&Element;{\mathrm{path}}_{sub}\end{array}---\left(5\right)$

Then, all water droplets are placed in connecting up starting point.

Path point selection, and record path nodal information are carried out using roulette method, formula is as follows：

$P_{ij}=\frac{f\left({\mathrm{soil}}_{ij}\right)}{{\mathrm{\&Sigma;}}_{j\&Element;N}f\left({\mathrm{soil}}_{ij}\right)}---\left(6\right)$

$f\left({\mathrm{soil}}_{ij}\right)=\left\{\begin{array}{cc}{\mathrm{soil}}_j\&times;\frac{a_f}{b_f+dis{({\mathrm{pos}}_i,{\mathrm{pos}}_j)}^3}& \begin{array}{cc}if& dis({\mathrm{pos}}_i,{\mathrm{pos}}_j)=eup\end{array}\\ {\mathrm{soil}}_j\&times;\frac{\mathrm{\&rho;}\&times;a_f}{b_f+dis{({\mathrm{pos}}_i,{\mathrm{pos}}_j)}^3}& otherwise\end{array}\right.---\left(7\right)$

When water droplet moves to another node j, the speed of water droplet is updated：

${\mathrm{vel}}_j={\mathrm{vel}}_i+\frac{{\mathrm{soil}}_j}{{\mathrm{soil}}_j+a_v}---\left(8\right)$

Meanwhile, update the soil of moved to node, and the soil that now water droplet is carried：

${\mathrm{soil}}_j={\mathrm{soil}}_j+{\mathrm{\&rho;}}_1\&times;\frac{{\mathrm{vel}}_j}{a_a+dis({\mathrm{pos}}_i,{\mathrm{pos}}_j)}---\left(9\right)$

${\mathrm{soil}}^{IWD}={\mathrm{soil}}^{IWD}+\frac{{\mathrm{vel}}_j}{a_s+dis({\mathrm{pos}}_i,{\mathrm{pos}}_j)}---\left(10\right)$

Before all water droplets move to wiring terminal, recurring formula (6) (7) (8) (9) (10) calculates each water droplet and is formed The fitness in path, and update locally optimal solution p_pWith globally optimal solution p_g；

After all water droplets move to wiring terminal, global amount of soil is updated：

${\mathrm{soil}}_i=\left\{\begin{array}{ccc}{\mathrm{soil}}_i+{\mathrm{\&rho;}}_2\&times;\frac{1}{num}{\mathrm{soil}}_{best}^{IWD}& if& {\mathrm{pos}}_i\&Element;p_p\\ {\mathrm{soil}}_i+{\mathrm{\&rho;}}_3\&times;\frac{1}{num}{\mathrm{soil}}_{best}^{IWD}& if& {\mathrm{pos}}_i\&Element;p_g\end{array}\right.---\left(11\right)$

After complete office's amount of soil is updated according to formula (11), algorithm completes an iteration.Hereafter algorithm will be into next time Iteration, all water droplets will be refitted in wiring starting point, and re-start movement to formula (11) with above-mentioned formula (6), until repeatedly Generation number reaches maximum or meets required precision.

6. it is according to claim 1 based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that step It is rapid 5) in, using the node in path and inflection point as control point, and according to step 1) in planar development mode by two dimension Control point coordinate transformation be wiring space in Three dimensions control point coordinates.

7. it is according to claim 1 based on the cable routing pathway searching method for mixing water droplet algorithm, it is characterised in that step It is rapid 6) in, using three rank B-spline path optimizings：

First obtaining step 5) in, control point p₀(x₀,y₀,z₀),...,p_n(x_n,y_n,z_n), determine the exponent number k=of B-spline curves 3, define the support Interval T=[t of SPL₀,t₁,...,t_n+k] such as following formula：

$t_i=\left\{\begin{array}{ccc}0& if& i<k+1\\ i-k& if& k+1\&le;i\&le;n\\ n-k+1& if& i>n\end{array}\right.---\left(12\right)$

Then the basic function N of SPL is calculated_i,k(t), following two formula：

$N_{i,1}\left(t\right)=\left\{\begin{array}{cc}1& \begin{array}{cc}if& t_i\&le;t\&le;t_{i+1}\end{array}\\ 0& otherwise\end{array}\right.---\left(13\right)$

$N_{i,k}=\frac{t-t_i}{t_{i+k-1}-t_i}\&times;N_{i,k-1}\left(t\right)+\frac{t_{i+k}-t}{t_{i+k}-t_{i+1}}\&times;N_{i+1,k-1}\left(t\right)---\left(14\right)$

The equation of three rank B-spline curves, such as following formula can finally be tried to achieve：

$P\left(t\right)=\underset{i=0}{\overset{n}{\&Sigma;}}{P}_i\&times;N_{i,k}\left(t\right)---\left(15\right)$