CN101667972B - Power communication network service routing method and device - Google Patents

Abstract

The invention discloses a power communication network service routing method and a device, and the method comprises the following steps: A. generating an object optimization function according to a network optimization object; B. determining an initial node of a service and taking the initial node as the current node; C utilizing the ant colony algorithm for calculating the transition probabilities of all chains which are connected with the current node, while the transition probabilities are obtained by the calculation according to information elements on all the current chains; D. selectingthe next jump node as the current node according to the transition probabilities; E. judging whether the selected next jump node is a destination node or not of the service or not; if so, taking the selected next jump node as the current node, determining the information elements on all the current chains according to the object optimization function, and then repeatedly implementing the step C to the step E; and otherwise, generating the routing which corresponds to the service according to the initial node of the service and all the selected next jump nodes. The method and the device can ensure the optimal performances of the multi-service routing in a power communication network.

Description

Power communication network service routing method and equipment

Technical field

The present invention relates to the powerline network technical field, be specifically related to a kind of power communication network service routing method and equipment.

Background technology

The planning of service path in the powerline network is meant under network environment and the definite condition of traffic matrix; with certain optimization aim; for each the professional evaluation work in the traffic matrix or protection route and reasonable disposition resource, satisfying the construction cost that minimizes network under the prerequisite of business demand.This problem itself is the multi-objective optimization question of a complexity, and the conventional method of head it off is generally integral linear programming method and some heuristics.Wherein integral linear programming method has higher time complexity, is therefore replaced by heuritic approach gradually in actual applications.

When the service path in the powerline network was planned, though can use for reference plan of operation method in the existing communication network, these methods all had its limitation.Such as, the integral linear programming method lacks the control to the search volume, so time complexity is higher relatively, and application difficult; And that genetic algorithm produces the similarity of separating between generations is higher, therefore is absorbed in locally optimal solution easily; The model of Artificial Immune Algorithm is simple, be easy to realize, but performance relative with portability a little less than; The ability of searching optimum of simulated annealing is stronger, and can jump out locally optimal solution effectively, but amount of calculation is bigger, the algorithm time complexity is higher, convergence rate is slow and have the problem of losing approximate optimal solution easily, therefore often can't satisfy the requirement that network need return many routes.

Summary of the invention

The embodiment of the invention provides a kind of power communication network service routing method and equipment, satisfying the parallel computation of multi-service route in the powerline network, and the best performance of the work of assurance route.

For this reason, the embodiment of the invention provides following technical scheme:

A kind of power communication network service routing method may further comprise the steps:

A, generate the objective optimization function according to network optimization target;

B, determine professional start node, and with it as present node;

The transition probability of C, each bar link of utilizing ant group algorithm to calculate to link to each other with present node, described transition probability is to calculate according to the pheromones on current each bar link;

D, the described transition probability of foundation are selected next-hop node;

Whether the next-hop node that E, judgement are selected is the destination node of described business; If, then with the next-hop node selected as present node, and determine pheromones on current each bar link according to described objective optimization function, repeated execution of steps C is to step e then; Otherwise, generate the route of corresponding described business according to the start node of business and all next-hop nodes of selecting.

A kind of power communication network service route device comprises:

The majorized function generation unit is used for generating the objective optimization function according to network optimization target;

The node selected cell is used for determining professional start node, and with its present node as ant group place;

The pheromones determining unit is used for determining pheromones on each bar link according to described objective optimization function;

The transition probability computing unit utilizes ant group algorithm to calculate the transition probability of each the bar link that links to each other with described present node, and described transition probability is to calculate according to the pheromones on current each bar link;

Described node selected cell also is used for determining next-hop node according to described transition probability, and with its present node as ant group place;

The route generation unit is used for generating according to the start node of business and all next-hop nodes of selecting the route of corresponding described business;

Judging unit is used to judge whether the next-hop node that described node selected cell is selected is the destination node of described business, if not, then notifies described pheromones determining unit to determine pheromones on current each bar link according to described objective optimization function; If then notify start node and all next-hop nodes of the selecting route that generate corresponding described business of described route generation unit according to business.

Power communication network service routing method that the embodiment of the invention provides and device are used for reference the search principle of separating of ant group algorithm, generate the objective optimization function according to network optimization target; Utilize in the ant group algorithm computing service routing procedure, determine pheromones on current each bar link according to described objective optimization function, according to the pheromones on current each bar link, the transition probability of each bar link that calculating links to each other with present node, select next-hop node according to described transition probability, realize the planning of multi-service route in the powerline network, and the best performance of the work of assurance route.

Description of drawings

Fig. 1 is the flow chart of embodiment of the invention power communication network service routing method;

Fig. 2 utilizes ant group algorithm to generate the flow chart of professional route in the embodiment of the invention;

Fig. 3 is a kind of structural representation of embodiment of the invention power communication network service route device.

Embodiment

In order to make those skilled in the art person understand the scheme of the embodiment of the invention better, the embodiment of the invention is described in further detail below in conjunction with drawings and embodiments.

At first the basic principle of ant group algorithm is done simple declaration below.

Ant group algorithm is emerging a kind of bionical evolution algorithm in the optimization field, and this algorithm adopts distributed parallel computer system, easily combines with additive method, has stronger robustness.Ant group algorithm comprises two root phases: laundering period and cooperation stage, in the laundering period, each candidate constantly adjusts self structure according to the information of accumulation; In the cooperation stage, by information interchange, better to separate between the candidate solution in the hope of producing performance, this is similar to the study mechanism of learning automaton.

The principle of ant group discovery shortest path and mechanism mainly are according to some local messages and utilize simple rule to make a strategic decision: at first, allow the ant can avoiding obstacles, secondly, allow ant find food, just need allow they travel through on the space have a few; Once more, if allow ant find the shortest path, need to calculate all possible path and their size of comparison so.Based on rule have:

1, scope: the observed scope of ant is a grid world, and it is speed radius (generally being 3) that ant has a parameter, and it can observed scope be exactly 3*3 the grid world so, and mobile distance is also within this scope.

2, environment: the environment at ant place is a virtual world, and barrier is wherein arranged, and other ant is arranged, also has pheromones, information have two kinds, and a kind of is to find the ant of the food food information element under spilling, and a kind of is the pheromones that finds the ant of the nest nest under spilling.Each ant all only can the interior environmental information of its scope of perception.Environment allows pheromones disappear with certain speed.

3, the rule of looking for food: can seek whether food is arranged in the scope of perception every ant, if having just directly in the past.Otherwise see if there is pheromones, and relatively in can the scope of perception which point pheromones maximum, like this, it is just walked towards the many places of pheromones, and how every ant can make mistakes with small probability, thereby is not that the past maximum point of pheromones moves.Ant looks for the rule of nest with top the same, and only it is made a response to the pheromones of nest, and to the plain no reaction of food information.

4, move rule: every ant all plain maximum direction of orientation information moves, and, when around when not having pheromones to guide, ant can go down according to the motion of own original travel direction inertia, and, a little disturbance is at random arranged in travel direction.In order to prevent that the ant original place from turn-taking, it can remember which point of just having passed by recently, if passing by recently more down of finding to walk, it will avoid as far as possible.

5, keep away the barrier rule: if the direction that ant will be moved has barrier to block, its meeting another direction of selection at random, and have pheromones to guide, its can be according to the regular behavior of looking for food.

6, sow the pheromones rule: every ant is spread the pheromones of sending out when just finding food or nest maximum, and along with it walks distance far away, the pheromones of sowing is fewer and feweri.

Embodiment of the invention power communication network service routing method is used for reference the search principle of separating of ant group algorithm, generates the objective optimization function according to network optimization target; Determine pheromones on each bar link according to described objective optimization function; Calculate the transition probability of each the bar link that links to each other with professional starting point according to described pheromones; And, realize the planning of multi-service route in the powerline network, and the best performance of the work of assurance route according to described transition probability selection next-hop node.

As shown in Figure 1, be the flow chart of embodiment of the invention power communication network service routing method, may further comprise the steps:

Step 101 generates the objective optimization function according to network optimization target.

Such as, be optimization aim with the economy, then when modeling (promptly generating majorized function), need consider network link cost and network node cost.Because it is very little that network operation under normal circumstances and wear and tear expense are compared with construction cost, generally can ignore in the planning stage, so the main construction cost of considering network in the model.Wherein, link cost is relevant with the cost of unit length optical fiber, and the node cost is relevant with the unit cost with the structure of node place switching equipment.

Suppose known network topology for directed graph G (V, F), wherein V and F represent the node set and the link set of network respectively, | F| represents the total network links number.W represents the set of single fiber wavelength, | W| represents single fiber wavelength sum.The total construction cost of C (G) expression the whole network, C _V(G) the total construction cost of expression the whole network node, C _F(G) the total construction cost of expression total network links, then the optimization aim function is minC (G), then:

C(G)＝C _F(G)+C _V(G) (1)

If $\∀f_{\mathrm{ij}}\∈F(i,j\∈V),$ D (f _Ij) expression optical fiber f _IjLength, α is the weight factor of link cost, then C _F(G) can be expressed as:

$C_F\left(G\right)=\mathrm{\α}\underset{f_{\mathrm{ij}}\∈F}{\mathrm{\Σ}}d\left(f_{\mathrm{ij}}\right)---\left(2\right)$

The total construction cost C of the whole network node _V(G) mainly comprise multiplexed/demodulation multiplexer cost, optical cross-matrix cost and wavelength shifter cost, β, γ and η represent the weight factor of above three costs respectively.C _V(G) can be expressed as:

$C_V\left(G\right)=C_V^{\mathrm{MUX}}\left(G\right)+C_V^{\mathrm{OXC}}\left(G\right)+C_V^{\mathrm{WC}}\left(G\right)---\left(3\right)$

Wherein, C _V ^MUX(G) expression node place multiplexed/total cost of demodulation multiplexer, multiplexed/demodulation multiplexer is divided into the multiplexed/demodulation multiplexer of road wavelength about local and non-this locality.The former number is relevant with professional number in the road up and down at local node place and single fiber maximum wavelength number, and the latter's number is relevant with the fiber port number at node place; C _V ^OXC(G) cost of expression node place optical cross-matrix; C _V ^WC(G) total cost of expression node place wavelength shifter.

If $\∀v\∈V,$ TU _vBeing illustrated in the professional number of setting out on a journey in node v place, promptly is the professional number of source node with node v; TD _vBeing illustrated in the professional number on road under the node v place, promptly is the professional number of destination node with node v, then C _V ^MUX(G) can be expressed as:

Suppose that the optical cross-matrix of intranodal is made up of the cascade of 2*2 optical switch, then node v place input/outbound port number is K _vOptical cross-connection equipment need by Individual 2*2 optical switch is formed, then C _V ^OXC(G) can be expressed as:

$C_V^{\mathrm{OXC}}\left(G\right)=\mathrm{\γ}\underset{v\∈V}{\mathrm{\Σ}}\left(\frac{K_v}{2}{\log }_2{K}_v\right)---\left(5\right)$

Suppose that network possesses the long ability to transform of all-wave, the number of node place wavelength shifter is relevant with straight-through professional number with this underground road.Use TH _vThe straight-through professional number in expression node v place, then C _Wc(G) can be expressed as:

$C_V^{\mathrm{WC}}\left(G\right)=\mathrm{\η}\underset{v\∈V}{\mathrm{\Σ}}({\mathrm{TD}}_v+{\mathrm{TH}}_v)---\left(6\right)$

Comprehensive above-mentioned every cost, the total construction cost function of the whole network is:

Promptly with formula (7) as majorized function.Certainly, the embodiment of the invention is not limited in above-mentioned majorized function, according to the difference of applied environment and purpose, can also generate other majorized functions according to actual needs.

Step 102 is determined professional start node, and with it as present node.

Step 103 utilizes ant group algorithm to calculate the transition probability of each the bar link that links to each other with present node, and described transition probability is to calculate according to the pheromones on current each bar link.

Step 104, the described transition probability of foundation are selected next-hop node.

Need avoid the generation of route ring when calculating route, therefore need set up effective taboo table update mechanism provides strictness for ant group's path point selection foundation for business.

The scale of supposing the ant group is m, f _Ij∈ F, i, j ∈ V, ant k (1≤k≤when m) selecting the direction of next jumping at the node i place, τ _k(i, j) (i j) goes up residual pheromone concentration to the expression link.δ _IjExpression link (i, visibility j), δ _Ij=1/d (f _Ij).λ ₁Relative importance (the λ of pheromone concentration during the expression routing ₁〉=0), λ ₂Relative importance (the λ of expression visibility ₂〉=0), definition p _k(i, j) for ant k transfers to the transition probability of j node by the i node, then:

$p_k(i,j)=\left\{\begin{array}{cc}\frac{{\mathrm{\τ}}_k^{{\mathrm{\λ}}_1}(i,j){\mathrm{\δ}}_{\mathrm{ij}}^{{\mathrm{\λ}}_2}}{\underset{r\∈{\mathrm{allowed}}_k}{\mathrm{\Σ}}{\mathrm{\τ}}_k^{{\mathrm{\λ}}_1}(i,r){\mathrm{\δ}}_{\mathrm{ir}}^{{\mathrm{\λ}}_2}}& j\∈{\mathrm{allowed}}_k\\ 0& \mathrm{else}\end{array}\right.---\left(\text{8}\right)$

Wherein, allowed _k=0,1,2 ..., | V|-1}-tabu _k, be the current node set that can select as next-hop node of ant k, promptly the also node set of process not of direct connected link mutually and ant, tabu are arranged with node i _kTaboo table for ant k.Initial time, the pheromone concentration on each paths equate, are constant C, and ant k selects next step route according to the transition probability on each bar adjacent link in motion process, and upgrade the taboo table.

Step 105 judges whether the next-hop node of selecting is the destination node of described business; If then execution in step 107; Otherwise execution in step 106.

Step 106 as present node, and is determined pheromones on current each bar link according to described objective optimization function with the next-hop node selected, returns step 103 then.

Step 107 generates the route of corresponding described business according to the start node of business and all next-hop nodes of selecting.

In embodiments of the present invention, ant group algorithm is used for the planning in power communication network service path, the no longer single length according to selected path of ant is upgraded the pheromone concentration on the respective link, upgrades pheromone concentration on the link but calculate the optimization aim that finishes the back network in road according to this.

Particularly, can decide the pheromones increment according to the optimization aim of current network (building the total cost value), and change, thereby influence the path point selection of follow-up ant process formula along with the variation of iterations such as the whole network.As time goes on, pheromones residual in the network can be volatilized gradually, the persistence of pheromones on parameter ρ (0≤ρ＜1) the expression network link, and then 1-ρ represents the disappearance degree of pheromone concentration.

In embodiments of the present invention, can utilize improved pheromones increment method of adjustment and/or pheromones update method to upgrade the link information element.Following three kinds of modes are specifically arranged:

1. the increment of computing information element at first, as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function; h _kBe the jumping figure of k the selected route of ant, f (h _k) be the weighting function of pheromones increment, be defined as follows:

$f\left(h_k\right)=\left\{\begin{array}{cc}\frac{h_D}{h_k}& h_k\≤h_{\max }\\ 0& h_k>h_{\max }\end{array}\right.,---\left(10\right)$

Wherein, h _DFor to adopt dijkstra's algorithm be the shortest path that calculates of current business by jumping figure, h _MaxBe predetermined threshold.

By weighted, less each the bar link that route comprised of jumping figure that is calculated by ant group algorithm will obtain more relatively pheromones gain; The pheromones that each bar link that route comprised that jumping figure is bigger increases is then less relatively.When the jumping figure of certain bar route surpasses threshold value h _MaxThe time, the pheromones increment of each bar link that this route comprised is zero, has controlled the length of the route that single ant process calculates so effectively.Hop count threshold value h _MaxValue can grid of reference scale, such as, can define h _MaxBe 3 times of shortest path jumping figure between current business sourcesink node, i.e. h _Max=3h _D

Then, adjust pheromone concentration on each bar link as follows:

τ′ _k(i，j)＝ρτ _k(i，j)+Δτ _k(i，j)，0≤ρ＜1， (11)

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link.

2. the increment of computing information element at first, as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function;

Then, adjust pheromone concentration on each bar link as follows:

${\mathrm{\&tau;}}_k^{\&prime;}(i,j)={\left[\mathrm{\&rho;}{\mathrm{\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_1}+{\left[\mathrm{\&Delta;}{\mathrm{\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_2},0\&le;\mathrm{\&rho;}<1,---\left(13\right)$

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link, ω ₁And ω ₂Be weighted factor.

3. according to the increment of above-mentioned formula (9) computing information element, and adjust pheromone concentration on each bar link according to above-mentioned formula (13).

Wherein, ω ₁Effect be the better solutions that finds in the bounding algorithm initial stage search stage comparatively blindly, make the pheromones increase on these better paths slower, prevent that algorithm is absorbed in locally optimal solution too early, has enlarged the hunting zone; And in the algorithm later stage, because the pace of change of power function is greater than the pace of change of linear function, better the pheromones pace of change on the path is also accelerated thereupon, thereby has accelerated the convergence rate in algorithm later stage.ω ₂Effect be the increasing degree of control single pheromones.As target function value C (G _k) during significant change, τ _k' (i, variation j) will be more obvious, thus the raising better solutions comprises the pheromones increment on the link, the dynamic modified solutions direction of search.If make ω ₁＞ω ₂, then algorithm is more paid close attention to ability of searching optimum, avoids converging to too early on the locally optimal solution; And if make ω ₁＜ω ₂, then algorithm is more paid close attention to convergence rate, can revise the direction of search and convergence in time when more excellent separating occurring faster.

Embodiment of the invention power communication network service routing method is used for reference the search principle of separating of ant group algorithm, generates the objective optimization function according to network optimization target; Determine pheromones on each bar link according to described objective optimization function; Calculate the transition probability of each the bar link that links to each other with professional starting point according to described pheromones; Utilize ant group algorithm to determine professional route according to described transition probability, realize the planning of professional route in the powerline network.

Further, the embodiment of the invention is improved the computational methods of pheromones increment and the update mechanism of pheromones, in the calculating of pheromones increment, consider the influence of hop count to network performance, adopted based on the pheromones increment of jumping figure and adjusted strategy, and, realized dynamic adjustment to the route solution space direction of search by the pheromones increment that every ant brought is carried out the jumping figure weighting; Simultaneously when upgrading the link information element, two have been introduced greater than 1 Weighted Index, make in initial operating stage search phase constraint better solutions comparatively blindly, prevent to converge to too early on the locally optimal solution, and expansion search volume, and, add rapid convergence in the amplitude of variation of moving later stage increasing pheromones increment, improve the search efficiency of understanding.

As shown in Figure 2, be to utilize ant group algorithm to generate the particular flow sheet of professional route in the embodiment of the invention.

Be that example describes with a business below, may further comprise the steps:

Step 201, input new business;

Step 202 is carried out initialization, comprising: arrangement network topology data, and initialization ant swarm parameter, it is A that ant group number is set _m, the number of every group of ant is A _n, (m=A _m* A _n), the professional chained list T of initialization _c, node chain Table V _cWith link chained list F _c, each chained list last bit is end with the sky; With link chained list F _cIn the plain concentration of initial information of all links be changed to constant C (C＞0), the plain increment of initialization information is 0, the taboo table tabu of each ant of initialization _kBe empty (1≤k≤m), the cyclic variable j of ant group is set _Ant=1, ant cyclic variable i in the group _Ant=1.

Step 203 places the source node s place of current business, wherein k=A with k ant _n* (j _Ant-1)+i _Ant

Step 204 for the transition probability of k ant by above-mentioned formula (8) calculating and present node adjacent link, and is selected next-hop node.

Step 205 judges that whether current transferable node set is empty, i.e. allowed whether _k=Null; If then execution in step 213; Otherwise execution in step 206.

Step 206 judges whether this node is professional destination node; If then execution in step 207; Otherwise execution in step 214.

Step 207, current ant k is that current business is selected a route, this route is carried out calculating the target function value of this moment after the wavelength resource preassignment, and calculate the pheromones increment of link that this route is passed through.

When distributing wavelength resource, can adopt initial hit strategy (First Fit), promptly, select first the idle wavelength in first optical fiber that has idle wavelength according to the order of optical fiber and wavelength numbering, distribute to business.

Step 208, the ant cyclic variable adds 1 in will organizing, i.e. i _Ant=i _Ant+ 1.

Step 209 judges that whether the ant cyclic variable promptly judges whether i greater than this group ant number in the group _Ant＞A _nIf then execution in step 210; Otherwise execution in step 215.

Step 210, the selected route of that ant of the pheromones increment maximum of bringing for network in this group is found out in the loop ends of this group ant, and upgrades this route according to formula (11) or formula (13) and comprise pheromone concentration on the link; And ant group cyclic variable added 1, even j _Ant=j _Ant+ 1, i _Ant=1.

Whether step 211 judges ant group cyclic variable greater than the ant group number, i.e. j whether _Ant＞A _mIf then execution in step 212; Otherwise execution in step 216.

Step 212, the ant loop ends determines that the route that converges to promptly is the current business final route, and utilizes the initial hit strategy to be current business route assignment wavelength.

When the route of determining to converge to, can stay the pheromones increment size separately by all ants that successfully calculate professional route relatively determines, particularly, the route that ant calculated that can select to stay pheromones increment maximum is the route of current business.

Step 213 is judged this ant death, empties taboo table tabu _k, go to step 208 then.

Step 214 adds the taboo table tabu that ant k carries with present node and selected next-hop node _kIn, ant k advances, and returns step 204 then.

Step 215 empties taboo table tabu _k, should the shared wavelength resource of business preassignment in the releasing network; Return step 203 then, begin the routing process of next ant.

Step 216 empties taboo table tabu _k, should the shared wavelength resource of business preassignment in the releasing network, return step 203 then, begin the circulation of next group ant.

Certainly,, can repeat above-mentioned flow process, be embodied as a plurality of professional professional routes that generate if the situation of a plurality of service concurrences is arranged.

As seen, the embodiment of the invention is used for reference the search principle of separating of ant group algorithm, and proposes improved pheromones update mechanism, and fast convergence rate is also controlled, and the tropism is strong for the solution space searcher, is applicable to the economy planning problem that solves power telecom network.This method has relatively stronger robustness, positive feedback and good distributed parallel computing capability simultaneously, can satisfy the parallel computation of many professional routes, and itself also has the high similarity with the network routing procedure, therefore is easy to transplant realize.

One of ordinary skill in the art will appreciate that all or part of step that realizes in the foregoing description method is to instruct relevant hardware to finish by program, described program can be stored in the computer read/write memory medium, described storage medium, as: ROM/RAM, magnetic disc, CD etc.

Correspondingly, the embodiment of the invention also provides a kind of power communication network service route device.As shown in Figure 3, be a kind of structural representation of this device.

In this embodiment, described device comprises: majorized function generation unit 301, node selected cell 302, pheromones determining unit 303, transition probability computing unit 304, route generation unit 305, judging unit 306.Wherein:

Majorized function generation unit 301 is used for generating the objective optimization function according to network optimization target;

Node selected cell 302 is used for determining professional start node, and with its present node as ant group place;

Pheromones determining unit 303 is used for determining pheromones on each bar link according to described objective optimization function;

Transition probability computing unit 304 utilizes ant group algorithm to calculate the transition probability of each the bar link that links to each other with described present node, and described transition probability is to calculate according to the pheromones on current each bar link;

Described node selected cell 302 also is used for determining next-hop node according to described transition probability, and with its present node as ant group place;

Route generation unit 305 is used for generating according to the start node of business and all next-hop nodes of selecting the route of corresponding described business;

Judging unit 306, be used to judge whether the next-hop node that described node selected cell 302 is selected is the destination node of described business, if not, then notify described pheromones determining unit 303 to determine pheromones on current each bar link according to described objective optimization function; If then notify start node and all next-hop nodes of the selecting route that generate corresponding described business of described route generation unit 305 according to business.

Wherein, majorized function generation unit 301 can be according to the path planning needs, can generate corresponding objective optimization function with different network optimization targets, such as, when being optimization aim with the economy, need need consider network link cost and network node cost when modeling (promptly generating majorized function), in this application, majorized function generation unit 301 specifically can generate objective optimization function: minC (G)=C as follows _F(G)+C _V(G), wherein, C _F(G) be the total construction cost of total network links, C _V(G) be the total construction cost of the whole network node.Description in detailed process such as the front embodiment of the invention power communication network service routing method does not repeat them here.

In embodiments of the present invention, described pheromones determining unit 303 can have multiple implementation, such as:

A kind of embodiment of described pheromones determining unit 303 comprises: the plain concentration of the first incremental computations subelement and the first information is adjusted subelement.Wherein:

The described first incremental computations subelement is used for the increment of computing information element as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function; h _kBe the jumping figure of k the selected route of ant, f (h _k) be the weighting function of pheromones increment, be defined as follows:

$f\left(h_k\right)=\left\{\begin{array}{cc}\frac{h_D}{h_k}& h_k\&le;h_{\max }\\ 0& h_k>h_{\max }\end{array}\right.,$

Wherein, h _DFor to adopt dijkstra's algorithm be the shortest path that calculates of current business by jumping figure, h _MaxBe predetermined threshold;

The plain concentration of the described first information is adjusted subelement, is used for adjusting as follows the pheromone concentration on each bar link:

τ′ _k(i，j)＝ρτ _k(i，j)+Δτ _k(i，j)，0≤ρ＜1，

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link.

A kind of embodiment of described pheromones determining unit 303 comprises: the second incremental computations subelement and second pheromone concentration are adjusted subelement.Wherein:

Described second increment is determined subelement, is used for the increment of computing information element as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function;

Described second pheromone concentration is adjusted subelement, is used for adjusting as follows the pheromone concentration on each bar link:

${\mathrm{\&tau;}}_k^{\&prime;}(i,j)=[\mathrm{\&rho;}{\mathrm{\&tau;}}_k{(i,j)]}^{{\mathrm{\&omega;}}_1}+{\left[\mathrm{\&Delta;}{\mathrm{\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_2},0\&le;\mathrm{\&rho;}<1,$

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link, ω ₁And ω ₂Be weighted factor.

Certainly, the another kind of embodiment of described pheromones determining unit 303 can also comprise: described first incremental computations subelement and described second pheromone concentration are adjusted subelement.

In embodiments of the present invention, described transition probability computing unit 304 specifically is used to utilize following formula to calculate the transition probability of each the bar link that links to each other with professional starting point:

$p_k(i,j)=\left\{\begin{array}{cc}\frac{{\mathrm{\&tau;}}_k^{{\mathrm{\&lambda;}}_1}(i,j){\mathrm{\&delta;}}_{\mathrm{ij}}^{{\mathrm{\&lambda;}}_2}}{\underset{r\&Element;{\mathrm{allowed}}_k}{\mathrm{\&Sigma;}}{\mathrm{\&tau;}}_k^{{\mathrm{\&lambda;}}_1}(i,r){\mathrm{\&delta;}}_{\mathrm{ir}}^{{\mathrm{\&lambda;}}_2}}& j\&Element;{\mathrm{allowed}}_k\\ 0& \mathrm{else}\end{array}\right.,$ λ ₁≥0，λ ₂≥0,

Wherein, p _k(i j) transfers to the transition probability of j node, δ for ant k by the i node _IjExpression link (i, visibility j), δ _Ij=1/d (f _Ij), λ ₁The relative importance of pheromone concentration during the expression routing, λ ₃The relative importance of expression visibility, allowed _k=0,1,2 ..., | V|-1}-tabu _k, for the also node set of process not of direct connected link mutually and ant, tabu being arranged with node i _kTaboo table for ant k.

The detailed process of utilizing embodiment of the invention power communication network service route device to generate professional route can not repeat them here with reference to the description in the embodiment of the invention power communication network service routing method of front.

Embodiment of the invention power communication network service route device is used for reference the search principle of separating of ant group algorithm, generates the objective optimization function according to network optimization target; Determine pheromones on each bar link according to described objective optimization function; Calculate the transition probability of each the bar link that links to each other with professional starting point according to described pheromones; And, realize the planning of multi-service route in the powerline network, and the best performance of the work of assurance route according to described transition probability selection next-hop node.

More than the embodiment of the invention is described in detail, used embodiment herein the present invention set forth, the explanation of above embodiment just is used for help understanding method and apparatus of the present invention; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, the part that all can change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Claims (10)

1. a power communication network service routing method is characterized in that, may further comprise the steps:

A, generate objective optimization function m inC (G)=C according to network optimization target _F(G)+C _V(G), wherein, C _F(G) be the total construction cost of total network links, C _V(G) be the total construction cost of the whole network node;

B, determine professional start node, and with it as present node;

The transition probability of C, each bar link of utilizing ant group algorithm to calculate to link to each other with present node, described transition probability is to calculate according to the pheromones on current each bar link;

D, the described transition probability of foundation are selected next-hop node;

Whether the next-hop node that E, judgement are selected is the destination node of described business; If, then with the next-hop node selected as present node, and determine pheromones on current each bar link according to described objective optimization function, repeated execution of steps C is to step e then; Otherwise, generate the route of corresponding described business according to the start node of business and all next-hop nodes of selecting.

2. method according to claim 1 is characterized in that, describedly determines that according to described objective optimization function the pheromones on current each bar link comprises:

The increment of computing information element as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function; h _kBe the jumping figure of k the selected route of ant, f (h _k) be the weighting function of pheromones increment, be defined as follows:

$f\left(h_k\right)=\left\{\begin{array}{cc}\frac{h_D}{h_k}& h_k\&le;h_{\max }\\ 0& h_k>h_{\max }\end{array}\right.,$

Wherein, h _DFor to adopt dijkstra's algorithm be the shortest path that calculates of current business by jumping figure, h _MaxBe predetermined threshold;

Adjust the pheromone concentration on each bar link as follows:

τ′ _k(i，j)＝ρτ _k(i，j)+Δτ _k(i，j)，0≤ρ＜1，

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link.

3. method according to claim 2 is characterized in that, described h _MaxBe 3 times of shortest path jumping figure between current business sourcesink node.

4. method according to claim 1 is characterized in that, describedly determines that according to described objective optimization function the pheromones on current each bar link comprises:

The increment of computing information element as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function;

Adjust the pheromone concentration on each bar link as follows:

${\mathrm{\&tau;}}_k^{\&prime;}(i,j)={\left[{\mathrm{\&rho;\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_1}+{\left[{\mathrm{\&Delta;\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_2},0\&le;\mathrm{\&rho;}<1,$

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link, ω ₁And ω ₂Be weighted factor.

5. method according to claim 4 is characterized in that, described method also comprises:

When needs are considered ability of searching optimum, set ω ₁＞ω ₂

When needs are considered convergence rate, set ω ₁＜ω ₂

6. according to claim 2 or 4 described methods, it is characterized in that the described transition probability that calculates each the bar link that links to each other with professional starting point according to described pheromones comprises:

Utilize following formula to calculate the transition probability of each the bar link that links to each other with professional starting point:

$p_k(i,j)=\left\{\begin{array}{cc}\frac{{\mathrm{\&tau;}}_k^{{\mathrm{\&lambda;}}_1}(i,j){\mathrm{\&delta;}}_{\mathrm{ij}}^{{\mathrm{\&lambda;}}_2}}{\underset{r\&Element;{\mathrm{allowed}}_k}{\mathrm{\&Sigma;}}{\mathrm{\&tau;}}_k^{{\mathrm{\&lambda;}}_1}(i,r){\mathrm{\&delta;}}_{\mathrm{ir}}^{{\mathrm{\&lambda;}}_2}}& j\&Element;{\mathrm{allowed}}_k\\ 0& \mathrm{else}\end{array}\right.,{\mathrm{\&lambda;}}_1\&GreaterEqual;0,{\mathrm{\&lambda;}}_2\&GreaterEqual;0,$

Wherein, p _k(i j) transfers to the transition probability of j node, δ for ant k by the i node _IjExpression link (i, visibility j), δ _Ij=1/d (f _Ij), λ ₁The relative importance of pheromone concentration during the expression routing, λ ₂The relative importance of expression visibility, allowed _k=0,1,2 ..., | V|-1}-tabu _k, for the also node set of process not of direct connected link mutually and ant, tabu being arranged with node i _kTaboo table for ant k.

7. a power communication network service route device is characterized in that, comprising:

The majorized function generation unit is used for generating objective optimization function m inC (G)=C according to network optimization target _F(G)+C _V(G), wherein, C _F(G) be the total construction cost of total network links, C _V(G) be the total construction cost of the whole network node;

The node selected cell is used for determining professional start node, and with its present node as ant group place;

The pheromones determining unit is used for determining pheromones on each bar link according to described objective optimization function;

The transition probability computing unit utilizes ant group algorithm to calculate the transition probability of each the bar link that links to each other with described present node, and described transition probability is to calculate according to the pheromones on current each bar link;

Described node selected cell also is used for determining next-hop node according to described transition probability, and with its present node as ant group place;

The route generation unit is used for generating according to the start node of business and all next-hop nodes of selecting the route of corresponding described business;

Judging unit is used to judge whether the next-hop node that described node selected cell is selected is the destination node of described business, if not, then notifies described pheromones determining unit to determine pheromones on current each bar link according to described objective optimization function; If then notify start node and all next-hop nodes of the selecting route that generate corresponding described business of described route generation unit according to business.

8. device according to claim 7 is characterized in that, described pheromones determining unit comprises:

The first incremental computations subelement is used for the increment of computing information element as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function; h _kBe the jumping figure of k the selected route of ant, f (h _k) be the weighting function of pheromones increment, be defined as follows:

$f\left(h_k\right)=\left\{\begin{array}{cc}\frac{h_D}{h_k}& h_k\&le;h_{\max }\\ 0& h_k>h_{\max }\end{array}\right.,$

Wherein, h _DFor to adopt dijkstra's algorithm be the shortest path that calculates of current business by jumping figure, h _MaxBe predetermined threshold;

The plain concentration of the first information is adjusted subelement, is used for adjusting as follows the pheromone concentration on each bar link:

τ′ _k(i，j)＝ρτ _k(i，j)+Δτ _k(i，j)，0≤ρ＜1，

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link.

9. device according to claim 7 is characterized in that, described pheromones determining unit comprises:

Second increment is determined subelement, is used for the increment of computing information element as follows:

Wherein, Δ τ _k(i, j) k ant of expression stayed link f in this circulation _IjOn the increment of pheromones; Q is one and embodies the constant that the unit ant stays track quantity; G _kRepresent that k ant is the network topology after current business calculates a route and Resources allocation, C (G _k) be current network topology G _kThe objective optimization function;

Second pheromone concentration is adjusted subelement, is used for adjusting as follows the pheromone concentration on each bar link:

${\mathrm{\&tau;}}_k^{\&prime;}(i,j)={\left[{\mathrm{\&rho;\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_1}+{\left[{\mathrm{\&Delta;\&tau;}}_k(i,j)\right]}^{{\mathrm{\&omega;}}_2},0\&le;\mathrm{\&rho;}<1,$

Wherein, τ _k(i, j) and τ ' _k(i j) represents before this pheromones renewal respectively and link f after upgrading _IjOn pheromone concentration, ρ represents the persistence of pheromones on the link, ω ₁And ω ₂Be weighted factor.

10. according to Claim 8 or 9 described devices, it is characterized in that,

Described transition probability computing unit specifically is used to utilize following formula to calculate the transition probability of each the bar link that links to each other with professional starting point:

$p_k(i,j)=\left\{\begin{array}{cc}\frac{{\mathrm{\&tau;}}_k^{{\mathrm{\&lambda;}}_1}(i,j){\mathrm{\&delta;}}_{\mathrm{ij}}^{{\mathrm{\&lambda;}}_2}}{\underset{r\&Element;{\mathrm{allowed}}_k}{\mathrm{\&Sigma;}}{\mathrm{\&tau;}}_k^{{\mathrm{\&lambda;}}_1}(i,r){\mathrm{\&delta;}}_{\mathrm{ir}}^{{\mathrm{\&lambda;}}_2}}& j\&Element;{\mathrm{allowed}}_k\\ 0& \mathrm{else}\end{array}\right.,{\mathrm{\&lambda;}}_1\&GreaterEqual;0,{\mathrm{\&lambda;}}_2\&GreaterEqual;0,$

Wherein, p _k(i j) transfers to the transition probability of j node, δ for ant k by the i node _IjExpression link (i, visibility j), δ _Ij=1/d (f _Ij), λ ₁The relative importance of pheromone concentration during the expression routing, λ ₂The relative importance of expression visibility, allowed _k=0,1,2 ..., | V|-1}-tabu _k, for the also node set of process not of direct connected link mutually and ant, tabu being arranged with node i _kTaboo table for ant k.

Images