CN102215168A - Method for optimizing and scheduling service resources based on laminated network - Google Patents

Abstract

The invention provides a method for realizing optimization and management of laminated network service flow through service distributing and scheduling the upper layer of the laminated network and routing and planning the lower layer of the laminated network. The method provided by the invention has the beneficial effects that a service scheduling method for satisfying the information transmission and service quality requirement in the laminated network is disclosed, through building an upper layer master problem optimization model and a lower layer sub problem optimization model of the laminated network, a feasible solution can be obtained, and the feasible region of the upper layer master problem optimization model can be reduced, thus the service resource optimization, scheduling and management capabilities in the laminated network can be improved.

Description

A kind of service resources optimized dispatching method based on stacked network

Technical field

The present invention relates to communication technical field and fuzzy control field, particularly relate to stacked network and service resources scheduling mechanism.

Background technology

In legacy network, the closer to the user, network demand is high more, and network is used the transmission clothes that provide to the user and is equipped with transparent more.This transparency simplified the user and used the complexity of network, but also lost the ability that the user controls network.Expectation has nothing in common with each other different application to network behavior, along with the appearance of the various application models of stacked network, and Development of Grid Technology, application that different demands stress and stacked network are urgent day by day to the requirement that participates in bearer network control.Stacked efficiently network needs lower floor's network that route controlled function and topology information are provided, and is optimized at physical network topology, as shown in Figure 1.

Service-oriented stacked network is made up by the carrying node, its major function is to make up a stacked network on legacy network and various dedicated network, the realization difference of shielding lower floor physical carrier network avoids various services in the traditional IP can only be deployed in the problem of network edge.Service-oriented cascade net ruton is crossed network aware and is guaranteed that whole stacked network has the topological structure of optimization, improves the adaptability of whole network system.Service-oriented stacked network model has extracted the basic general character of various application services in the network, and they are reduced the new network agreement based on cascade net such as application layer multicast, large-scale stream media, application layer QoS.In the stacked pessimistic concurrency control of many services, with the new network agreement of these the whole network scales service content as the infrastructure of whole network, directly provide the network service of enhancing by many services cascade net, solved the problem that traditional IP can't adapt to all application and service demands by stacked network technology based on application layer to the upper strata.There is polynary isomery characteristic in Business Stream in the stacked network, therefore is necessary the service resources in the stacked network is managed efficiently and dispatched.

In sum: be necessary in stacked network, to adopt the dispatching method of efficient layering, realize service resources optimized dispatching and management.

Summary of the invention

Technical problem to be solved by this invention is: realize the service resources optimized dispatching in the stacked network.

The present invention solves the problems of the technologies described above the technical scheme that is adopted to be: by stacked upper network layer being carried out traffic assignments and scheduling and lower floor being carried out the optimum management of route planning realization to stacked network service flow; This method is by setting up stacked upper network layer primal problem Optimization Model and lower floor's subproblem Optimization Model, and the feasible zone that obtains feasible solution and reduce upper strata primal problem Optimization Model, thereby improves service resources optimized dispatching and managerial ability in the stacked network; It is characterized in that: improve service resource management ability in the transmission course by inventing method that stacked upper network layer primal problem Optimization Model and lower floor's subproblem Optimization Model combine, may further comprise the steps:

A, set up stacked upper network layer primal problem Optimization Model;

B, set up stacked network lower floor subproblem Optimization Model;

C, structure also obtain steps A and the feasible solution of the Optimization Model that B is set up;

The feasible zone of D, minimizing upper strata primal problem Optimization Model.

In the described steps A, set up stacked upper network layer primal problem Optimization Model, this model is used for service dispatching and transmission link distributes, and the link capacity constraints in the transmission course at one time is ${\mathrm{\θ}}_{j,l}-t_{j,l}^u+S(1-z_{i,j,u}^l)+S(2-{\mathrm{\γ}}_{i,u}^l-{\mathrm{\γ}}_{j,u}^l)\≥{\mathrm{\θ}}_{i,l}$ Know ${\mathrm{\θ}}_{i,l}-t_{i,l}^u+{\mathrm{Sz}}_{i,j,u}^l+S(2-{\mathrm{\γ}}_{i,u}^l-{\mathrm{\γ}}_{j,u}^l)\≥{\mathrm{\θ}}_{i,l},,$ Wherein

$\∀i,j\∈J|i\≠j,\∀u\∈M_{j,l},\∀l\∈P\∪P_{\mathrm{AGV}}$ $\∀i,j\∈J|i\≠j,\∀u\∈M_{i,l},\∀l\∈P\∪P_{\mathrm{AGV}}$ With

Be constant, θ _{I, l}Be the operation deadline of Business Stream i at transmitting scene l, For in the operational processes time of transmitting scene l and node u Business Stream i,

Be binary variable, if in time τ, Business Stream k is from node n _iTo n _j, then

Otherwise,

Be binary variable, if carry out the operation of Business Stream i, then at transmitting scene l

Otherwise,

Be binary variable, if at node u, transmitting scene l, Business Stream i is processed prior to Business Stream j, then

Otherwise,

Therefore upper strata primal problem Optimization Model can be described as $\min \underset{i\∈J}{\mathrm{\Σ}}{w}_i{T}_i,$ $T_i=\max \{0,(c_{i,o}-D_i)\}$

$s.t.z_{i,j,u}^l\≤{\mathrm{\γ}}_{i,u}^l,\∀u\∈M_{i,l},\∀i,j\∈J|i\≠j,\∀l\∈P\∪P_{\mathrm{AGV}},---\left(1\right)$

$z_{i,j,u}^l\≤{\mathrm{\γ}}_{j,u}^l,\∀u\∈M_{j,l},\∀i,j\∈J|i\≠j,\∀l\∈P\∪P_{\mathrm{AGV}},---\left(2\right)$

$\underset{u\∈M_{i,j}}{\mathrm{\Σ}}{\mathrm{\γ}}_{i,u}^l=1,\∀i\∈J,\∀l\∈P\∪P_{\mathrm{AGV}},---\left(3\right)$

${\mathrm{\θ}}_{j,l}-t_{j,l}^u+S(1-z_{i,j,u}^l)+S(2-{\mathrm{\γ}}_{i,u}^l-{\mathrm{\γ}}_{j,u}^l)\≥{\mathrm{\θ}}_{i,l},$ (4)

$\∀i,j\∈J|i\≠j,\∀u\∈M_{j,l},\∀l\∈P\∪P_{\mathrm{AGV}},$

${\mathrm{\θ}}_{i,l}-t_{i,l}^u+{\mathrm{Sz}}_{i,j,u}^l+S(2-{\mathrm{\γ}}_{i,u}^l-{\mathrm{\γ}}_{j,u}^l)\≥{\mathrm{\θ}}_{j,l},$ (5)

$\∀i,j\∈J|i\≠j,\∀u\∈M_{i,l},\∀l\∈P\∪P_{\mathrm{AGV}}$

${\mathrm{\θ}}_{j,p\left(l\right)}-t_{j,p\left(l\right)}^u\≥{\mathrm{\θ}}_{j,l},\∀j\∈J,\∀u\∈M_{j,l},\∀l\∈P\∪P_{\mathrm{AGV}},---\left(6\right)$

$\underset{n_j\∉E_{n_i}}{\mathrm{\Σ}}{\mathrm{\η}}_{n_i,n_j,\mathrm{\τ}}^k=0,\∀k\∈V,\∀n_i\∈E,1\≤\mathrm{\τ}\≤H,---\left(7\right)$

$\underset{n_j\∈E_{n_i}}{\mathrm{\Σ}}{\mathrm{\η}}_{n_i,n_j,\mathrm{\τ}}^k\≤1,\∀k\∈V,{\∀n}_i\∈E,1\≤\mathrm{\τ}\≤H,---\left(8\right)$

$\underset{n_j\∈E_{n_i}}{\mathrm{\Σ}}{\mathrm{\η}}_{n_j,n_i,\mathrm{\τ}}^k=\underset{n_l\∈E_{n_i}}{\mathrm{\Σ}}{\mathrm{\η}}_{n_i,n_j,\mathrm{\τ}+1,}^k\∀k\∈V,{\∀n}_i\∈E,1\≤\mathrm{\τ}\≤H-1,---\left(9\right)$

$\underset{n_j\∈E_{v_k}}{\mathrm{\Σ}}{\mathrm{\η}}_{v_k,n_j,1}^k=1,\∀k\∈V,---\left(10\right)$

$\underset{k\∈V}{\mathrm{\Σ}}\underset{n_j\∈E}{\mathrm{\Σ}}{\mathrm{\η}}_{n_j{n}_i,\mathrm{\τ}}^k\≤1,\∀n_i\∈E,1\≤\mathrm{\τ}\≤H,---\left(11\right)$

$\underset{k\∈V}{\mathrm{\Σ}}({\mathrm{\η}}_{n_i,n_j,\mathrm{\τ}}^k+{\mathrm{\η}}_{n_j,n_i,\mathrm{\τ}}^k)\≤1,\∀n_i\∈E,\∀n_j\∈E_{n_i},1\≤\mathrm{\τ}\≤H,---\left(12\right)$

$\underset{n_j\∈E_{s_i^l}}{\mathrm{\Σ}}{\mathrm{\η}}_{s_i^l,n_j,\mathrm{\τ}+1}^k\≥{\mathrm{\α}}_{i,\mathrm{\τ},k,}^l\∀k\∈V,\∀l\∈P_{\mathrm{AGV}},1\≤\mathrm{\τ}\≤H-1,---\left(13\right)$

$\underset{n_j\∈E_{F_i^l}}{\mathrm{\Σ}}{\mathrm{\η}}_{n_j,F_i^l,\mathrm{\τ}}^k\≥{\mathrm{\β}}_{i,\mathrm{\τ},k}^l,\∀k\∈V,\∀l\∈P_{\mathrm{AGV}},1\≤\mathrm{\τ}\≤H,---\left(14\right)$

${\mathrm{\θ}}_{i,l}-t_{i,l}+1\≤\mathrm{\τ}+S(1-{\mathrm{\μ}}_{i,\mathrm{\τ}}^l),---\left(15\right)$

${\mathrm{\θ}}_{i,l}-t_{i,l}+1>\mathrm{\τ}-{\mathrm{S\μ}}_{i,\mathrm{\τ}}^l,\∀i\∈J,1\≤\mathrm{\τ}\≤H,\∀l\∈P_{\mathrm{AGV}},---\left(16\right)$

${\mathrm{\θ}}_{i,l}+S(1-{\stackrel{\‾}{\mathrm{\μ}}}_{i,\mathrm{\τ}}^l)\≥\mathrm{\τ},---\left(17\right)$

${\mathrm{\&theta;}}_{i,l}<\mathrm{\&tau;}+S{\stackrel{-}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;l\&Element;P_{\mathrm{AGV}},---\left(18\right)$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&alpha;}}_{i,\mathrm{\&tau;},u}^l)\&GreaterEqual;2,$ (19)

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;1+{\mathrm{S\&alpha;}}_{i,t,u}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},$

${\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&beta;}}_{i,\mathrm{\&tau;},u}^l)\&GreaterEqual;2,$ (20)

${\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;1+S{\mathrm{\&beta;}}_{i,t,u}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&sigma;}}_{i,\mathrm{\&tau;},k}^l)\&GreaterEqual;3,---\left(21\right)$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;2+{\mathrm{S\&sigma;}}_{i,\mathrm{\&tau;},k,}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},\left(22\right).$

Among the described step B, set up stacked network lower floor subproblem Optimization Model, be used to carry out the selection in path, this step is satisfying under the condition of steps A, judges whether the non-conflict route of multi-path transmission exists.Exist if satisfy the route of above-mentioned condition A, then go to step D.In stacked network, do not exist if satisfy the route of above-mentioned condition A, then adopt distributed routing algorithm, obtain the optimum transmission time and postpone punishment.The zero-time and the deadline of professional transmission are set then,

And σ _{I, τ, u}, in satisfy condition (19)-(23) afterwards, the solving-optimizing model: S.t. (7)-(14).In stacked network, increase traffic flow unit gradually, by in remaining Business Stream, inserting first traffic flow unit, produce k candidate sequence, in this sequence for current separating, select the sequential value of overall deadline local optimum, upgrade selecteed locally optimal solution and separate as current.If k=n then can obtain optimal solution, otherwise, then go to step C, as shown in Figure 2.

Among the described step C,, then need the feasible solution of the Optimization Model that construction step A and B set up in this step if the Optimization Model that steps A and B are set up has infeasible solution.Substep a. finds the solution stacked network lower floor subproblem, calculate nearest professional propagation delay time, if constraints does not exist, service condition H (professional arrive time lag by the start node of separating acquisition of stacked network lower floor subproblem) then in the transmission start time of separating acquisition by stacked upper network layer primal problem, enter substep b, otherwise, constraints does not exist, then service condition I (professional arrive time lag of separating terminal point by stacked network lower floor subproblem in being separated by the optimum operation of separating acquisition of upper strata primal problem) enters substep c.Substep b. postpones to handle to Business Stream ω makes the constraints of other Business Stream be satisfied, utilization B _{ω, ω+1}=(θ _ω+1-t _ω+1)-(θ _ω+ R (F _{ω, j}, S _{ω+1, l})) calculate time of delay of next Business Stream ω+1, wherein R (F _{ω, j}, S _{ω+1, l}) be the receiving node F of Business Stream ω _{ω, l}Start node S with Business Stream ω+1 _{ω+1, l}Between minimum range, if feasible solution exists, then stop to calculate, otherwise, then be back to substep a, delay if possible is 0, then goes to substep d.Making the required minimum growth time delay of the transmission deadline of Business Stream ω is d _ω, B _{ω, ω+1}With

In minimum value.The use last time ${\mathrm{\&theta;}}_{\mathrm{\&omega;}}+\min \{d_{\mathrm{\&omega;}},\min \{d_{\mathrm{\&omega;}},\min \{({\mathrm{\&theta;}}_{\mathrm{\&omega;}}^s-t_{\mathrm{\&omega;}}^s)-{\mathrm{\&theta;}}_{\mathrm{\&omega;}},B_{\mathrm{\&omega;}+1}\}\left\}\right\}$ The more transmission deadline of new service flow ω, and use $t_{\mathrm{\&omega;}}+\min \{d_{\mathrm{\&omega;}},\min \{d_{\mathrm{\&omega;}},\min \{({\mathrm{\&theta;}}_{\mathrm{\&omega;}}^s-t_{\mathrm{\&omega;}}^s)-{\mathrm{\&theta;}}_{\mathrm{\&omega;}},B_{\mathrm{\&omega;}+1}\}\left\}\right\}$ More the transmission time of new service flow ω, find the solution for lower floor's subproblem Optimization Model, if feasible solution exists, then algorithm stops, otherwise is back to substep a, and be 0 time of delay if possible, then goes to substep d.Substep c, more the transmission deadline of new service flow ω.Substep d, the orlop subproblem is found the solution, and d time of delay of computing service stream ω _ωSubstep e uses the transmission deadline θ of last Business Stream ω _ω+ d _ωAs its transmission deadline θ next time _ωUse this θ _α+ d _αAs business transmission deadline θ next time _α, wherein $d_{\mathrm{\&alpha;}}=\left\{\begin{array}{cc}{d}_{\mathrm{\&alpha;}-1}& (B_{\mathrm{\&alpha;}-1,\mathrm{\&alpha;}}=0)\\ \max \left(d_{\mathrm{\&alpha;}-1}{-B}_{\mathrm{\&alpha;}-\mathrm{1,}\mathrm{\&alpha;},0}\right)& (B_{\mathrm{\&alpha;}-1,\mathrm{\&alpha;}}\&NotEqual;0)\end{array}\right.,$ $d_{{\mathrm{\&alpha;}}^{\&prime;}}^s=\max \{0,d_e-\min \{({\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}^s-t_{{\mathrm{\&alpha;}}^{\&prime;}}^s)+d_e-{\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}-1,}^s({\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}^s-t_{{\mathrm{\&alpha;}}^{\&prime;}}^s)+d_e-{\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}\left\}\right\}$ For satisfying the minimum delay of sequence constraints.Lower floor's subproblem Optimization Model is found the solution,, then stopped if can obtaining feasible solution, otherwise, go to steps A.Substep f uses last deadline and transmission time θ _ω+ d _ωAnd t _ω+ d _ωAs deadline and the transmission time of Business Stream e next time, the lower floor subproblem is found the solution, if feasible solution exists, then stop, otherwise, then be back to substep a.

Among the described step C, convergence domain is assessed, the difference of obtaining the upper limit and the lower limit that obtains from steps A from step B and C is ε, and enters next step.

Among the described step D, get rid of previously obtd possible arrangement by using the integer cutting method, this possible arrangement can reduce the feasible zone of upper strata primal problem Optimization Model.

Be separating of the r time iteration acquisition.For making primal problem obtain to dispose at the difference of variable, use the integer cutting method in the upper strata subproblem,

$\underset{(i,u,l)\&Element;Y^1\left(r\right)}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l-\underset{(i,u,l)\&Element;Y^0\left(r\right)}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l\&le;\left|Y^1\left(r\right)\right|,\&ForAll;r,$

$\underset{(i,j,u,l)\&Element;z^1\left(r\right)}{\mathrm{\&Sigma;}}{z}_{i,j,u}^l-\underset{(i,j,u,l)\&Element;z^0\left(r\right)}{\mathrm{\&Sigma;}}{z}_{i,j,u}^l\&le;\left|z^1\left(r\right)\right|,\&ForAll;r,$

$Y^0\left(r\right)=\left\{(i,u,l)\right|{\mathrm{\&gamma;}}_{i,u}^l\left(r\right)=0\},$ $Y^1\left(r\right)=\left\{(i,u,l)\right|{\mathrm{\&gamma;}}_{i,u}^l\left(r\right)=1\},,$

$Z^0\left(r\right)=\left\{(i,j,u,l)\right|z_{i,j,u}^l\left(r\right)=0\},$ $Z^1\left(r\right)=\left\{(i,j,u,l)\right|z_{i,,j,u}^l\left(r\right)=1\}$

Wherein

Be the variable in the iteration people, by separating of from iteration, obtaining, the feasible zone of restriction upper strata primal problem.

Beneficial effect of the present invention is: proposed a kind of service resources optimized dispatching method based on stacked network, by realizing hierarchy optimization and mutually combining, effectively improved efficiency of transmission professional in the stacked network.

Description of drawings

Fig. 1 is stacked schematic network structure

Fig. 2 is the service resources optimized dispatching schematic flow sheet of stacked network

Embodiment

With embodiment the present invention is described in further detail with reference to the accompanying drawings below:

Basic ideas of the present invention are as follows: by stacked upper network layer being carried out traffic assignments and scheduling and lower floor being carried out the optimum management of route planning realization to stacked network service flow, realize optimization transmission professional in the stacked network.

1. set up stacked upper network layer primal problem Optimization Model, this model is used for service dispatching and transmission link distributes, and the link capacity constraints in the transmission course at one time is ${\mathrm{\&theta;}}_{j,l}-t_{j,l}^u+S(1-z_{i,j,u}^l)+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{i,l}$ With ${\mathrm{\&theta;}}_{i,l}-t_{i,l}^u+{\mathrm{Sz}}_{i,j,u}^l+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{i,l},,$ Wherein

$\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{j,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}}$ $\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}}$ Know Be constant, θ _{I, l}Be the operation deadline of Business Stream i at transmitting scene l, For in the operational processes time of transmitting scene l and node u Business Stream i,

Be binary variable, if in time τ, Business Stream k is from node n _iTo n _j, then

Otherwise,

Be binary variable, if carry out the operation of Business Stream i, then at transmitting scene l

Otherwise,

Be binary variable, if at node u, transmitting scene l, Business Stream i is processed prior to Business Stream j, then

Otherwise,

Therefore upper strata primal problem Optimization Model can be described as $\min \underset{i\&Element;J}{\mathrm{\&Sigma;}}{w}_i{T}_i,$ $T_i=\max \{0,(c_{i,o}-D_i)\}$

$s.t.z_{i,j,u}^l\&le;{\mathrm{\&gamma;}}_{i,u}^l,\&ForAll;u\&Element;M_{i,l},\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(1\right)$

$z_{i,j,u}^l\&le;{\mathrm{\&gamma;}}_{j,u}^l,\&ForAll;u\&Element;M_{j,l},\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(2\right)$

$\underset{u\&Element;M_{i,j}}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l=1,\&ForAll;i\&Element;J,\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(3\right)$

${\mathrm{\&theta;}}_{j,l}-t_{j,l}^u+S(1-z_{i,j,u}^l)+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{i,l},$ (4)

$\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{j,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},$

${\mathrm{\&theta;}}_{i,l}-t_{i,l}^u+{\mathrm{Sz}}_{i,j,u}^l+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{j,l},$ (5)

$\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}}$

${\mathrm{\&theta;}}_{j,p\left(l\right)}-t_{j,p\left(l\right)}^u\&GreaterEqual;{\mathrm{\&theta;}}_{j,l},\&ForAll;j\&Element;J,\&ForAll;u\&Element;M_{j,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(6\right)$

$\underset{n_j\&NotElement;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}}^k=0,\&ForAll;k\&Element;V,\&ForAll;n_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H,---\left(7\right)$

$\underset{n_j\&Element;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}}^k\&le;1,\&ForAll;k\&Element;V,{\&ForAll;n}_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H,---\left(8\right)$

$\underset{n_j\&Element;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_j,n_i,\mathrm{\&tau;}}^k=\underset{n_l\&Element;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}+1,}^k\&ForAll;k\&Element;V,{\&ForAll;n}_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H-1,---\left(9\right)$

$\underset{n_j\&Element;E_{v_k}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{v_k,n_j,1}^k=1,\&ForAll;k\&Element;V,---\left(10\right)$

$\underset{k\&Element;V}{\mathrm{\&Sigma;}}\underset{n_j\&Element;E}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_j{n}_i,\mathrm{\&tau;}}^k\&le;1,\&ForAll;n_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H,---\left(11\right)$

$\underset{k\&Element;V}{\mathrm{\&Sigma;}}({\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}}^k+{\mathrm{\&eta;}}_{n_j,n_i,\mathrm{\&tau;}}^k)\&le;1,\&ForAll;n_i\&Element;E,\&ForAll;n_j\&Element;E_{n_i},1\&le;\mathrm{\&tau;}\&le;H,---\left(12\right)$

$\underset{n_j\&Element;E_{s_i^l}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{s_i^l,n_j,\mathrm{\&tau;}+1}^k\&GreaterEqual;{\mathrm{\&alpha;}}_{i,\mathrm{\&tau;},k,}^l\&ForAll;k\&Element;V,\&ForAll;l\&Element;P_{\mathrm{AGV}},1\&le;\mathrm{\&tau;}\&le;H-1,---\left(13\right)$

$\underset{n_j\&Element;E_{F_i^l}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_j,F_i^l,\mathrm{\&tau;}}^k\&GreaterEqual;{\mathrm{\&beta;}}_{i,\mathrm{\&tau;},k}^l,\&ForAll;k\&Element;V,\&ForAll;l\&Element;P_{\mathrm{AGV}},1\&le;\mathrm{\&tau;}\&le;H,---\left(14\right)$

${\mathrm{\&theta;}}_{i,l}-t_{i,l}+1\&le;\mathrm{\&tau;}+S(1-{\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l),---\left(15\right)$

${\mathrm{\&theta;}}_{i,l}-t_{i,l}+1>\mathrm{\&tau;}-{\mathrm{S\&mu;}}_{i,\mathrm{\&tau;}}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;l\&Element;P_{\mathrm{AGV}},---\left(16\right)$

${\mathrm{\&theta;}}_{i,l}+S(1-{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l)\&GreaterEqual;\mathrm{\&tau;},---\left(17\right)$

${\mathrm{\&theta;}}_{i,l}<\mathrm{\&tau;}+S{\stackrel{-}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;l\&Element;P_{\mathrm{AGV}},---\left(18\right)$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&alpha;}}_{i,\mathrm{\&tau;},u}^l)\&GreaterEqual;2,$ (19)

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;1+{\mathrm{S\&alpha;}}_{i,t,u}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},$

${\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&beta;}}_{i,\mathrm{\&tau;},u}^l)\&GreaterEqual;2,$ (20)

${\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;1+S{\mathrm{\&beta;}}_{i,t,u}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&sigma;}}_{i,\mathrm{\&tau;},k}^l)\&GreaterEqual;3,---\left(21\right)$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;2+{\mathrm{S\&sigma;}}_{i,\mathrm{\&tau;},k,}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},\left(22\right).$

2. set up stacked network lower floor subproblem Optimization Model, be used to carry out the selection in path, this step is satisfying under the condition of step 1, judges whether the non-conflict route of multi-path transmission exists.Exist if satisfy the route of above-mentioned condition A, then go to step 4.In stacked network, do not exist if satisfy the route of above-mentioned condition 1, then adopt distributed routing algorithm, obtain the optimum transmission time and postpone punishment.The zero-time and the deadline of professional transmission are set then,

And σ _{I, τ, u}, in satisfy condition (19)-(23) afterwards, the solving-optimizing model:

S.t.(7)-(14)。In stacked network, increase traffic flow unit gradually, by in remaining Business Stream, inserting first traffic flow unit, produce k candidate sequence, in this sequence for current separating, select the sequential value of overall deadline local optimum, upgrade selecteed locally optimal solution and separate as current.If k=n then can obtain optimal solution, otherwise, then go to step 3, as shown in Figure 2.

3., then need the feasible solution of construction step 1 and 2 Optimization Model of being set up in this step if step 1 and 2 Optimization Model of being set up have infeasible solution.Substep a. finds the solution stacked network lower floor subproblem, calculate nearest professional propagation delay time, if constraints does not exist, service condition H (professional arrive time lag by the start node of separating acquisition of stacked network lower floor subproblem) then in the transmission start time of separating acquisition by stacked upper network layer primal problem, enter substep b, otherwise, constraints does not exist, then service condition I (professional arrive time lag of separating terminal point by stacked network lower floor subproblem in being separated by the optimum operation of separating acquisition of upper strata primal problem) enters substep c.Substep b. postpones to handle to Business Stream ω makes the constraints of other Business Stream be satisfied, utilization B _{ω, ω+1}=(θ _ω+1-t _ω+1)-(θ _ω+ R (F _{ω, j}, S _{ω+1, l})) calculate time of delay of next Business Stream ω+1, wherein R (F _{ω, j}, S _{ω+1, l}) be the receiving node F of Business Stream ω _{ω, l}Start node S with Business Stream ω+1 _{ω+1, l}Between minimum range, if feasible solution exists, then stop to calculate, otherwise, then be back to substep a, delay if possible is 0, then goes to substep d.Making the required minimum growth time delay of the transmission deadline of Business Stream ω is d _ω, B _{ω, ω+1}With

In minimum value.The use last time ${\mathrm{\&theta;}}_{\mathrm{\&omega;}}+\min \{d_{\mathrm{\&omega;}},\min \{d_{\mathrm{\&omega;}},\min \{({\mathrm{\&theta;}}_{\mathrm{\&omega;}}^s-t_{\mathrm{\&omega;}}^s)-{\mathrm{\&theta;}}_{\mathrm{\&omega;}},B_{\mathrm{\&omega;}+1}\}\left\}\right\}$ The more transmission deadline of new service flow ω, and use $t_{\mathrm{\&omega;}}+\min \{d_{\mathrm{\&omega;}},\min \{d_{\mathrm{\&omega;}},\min \{({\mathrm{\&theta;}}_{\mathrm{\&omega;}}^s-t_{\mathrm{\&omega;}}^s)-{\mathrm{\&theta;}}_{\mathrm{\&omega;}},B_{\mathrm{\&omega;}+1}\}\left\}\right\}$ More the transmission time of new service flow ω, find the solution for lower floor's subproblem Optimization Model, if feasible solution exists, then algorithm stops, otherwise is back to substep a, and be 0 time of delay if possible, then goes to substep d.Substep c, more the transmission deadline of new service flow ω.Substep d, the orlop subproblem is found the solution, and d time of delay of computing service stream ω _ωSubstep e uses the transmission deadline θ of last Business Stream ω _ω+ d _ωAs its transmission deadline θ next time _ωUse this θ _α+ d _αAs business transmission deadline θ next time _α, wherein $d_{\mathrm{\&alpha;}}=\left\{\begin{array}{cc}{d}_{\mathrm{\&alpha;}-1}& (B_{\mathrm{\&alpha;}-1,\mathrm{\&alpha;}}=0)\\ \max \left(d_{\mathrm{\&alpha;}-1}{-B}_{\mathrm{\&alpha;}-\mathrm{1,}\mathrm{\&alpha;},0}\right)& (B_{\mathrm{\&alpha;}-1,\mathrm{\&alpha;}}\&NotEqual;0)\end{array}\right.,$ $d_{{\mathrm{\&alpha;}}^{\&prime;}}^s=\max \{0,d_e-\min \{({\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}^s-t_{{\mathrm{\&alpha;}}^{\&prime;}}^s)+d_e-{\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}-1,}^s({\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}^s-t_{{\mathrm{\&alpha;}}^{\&prime;}}^s)+d_e-{\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}\left\}\right\}$ For satisfying the minimum delay of sequence constraints.Lower floor's subproblem Optimization Model is found the solution,, then stopped if can obtaining feasible solution, otherwise, go to step 1.Substep f uses last deadline and transmission time θ _ω+ d _ωAnd t _ω+ d _ωAs deadline and the transmission time of Business Stream e next time, the lower floor subproblem is found the solution, if feasible solution exists, then stop, otherwise, then be back to substep a.

4. convergence domain is assessed, the difference of obtaining the upper limit and the lower limit that obtains from step 1 from step 2 and 3 is ε, and enters next step.

5. by using the integer cutting method to get rid of previously obtd possible arrangement, this possible arrangement can reduce the feasible zone of upper strata primal problem Optimization Model. Be separating of the r time iteration acquisition.For making primal problem obtain to dispose at the difference of variable, use the integer cutting method in the upper strata subproblem,

$\underset{(i,u,l)\&Element;Y^1\left(r\right)}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l-\underset{(i,u,l)\&Element;Y^0\left(r\right)}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l\&le;\left|Y^1\left(r\right)\right|,\&ForAll;r,$

$\underset{(i,j,u,l)\&Element;z^1\left(r\right)}{\mathrm{\&Sigma;}}{z}_{i,j,u}^l-\underset{(i,j,u,l)\&Element;z^0\left(r\right)}{\mathrm{\&Sigma;}}{z}_{i,j,u}^l\&le;\left|z^1\left(r\right)\right|,\&ForAll;r,$

$Y^0\left(r\right)=\left\{(i,u,l)\right|{\mathrm{\&gamma;}}_{i,u}^l\left(r\right)=0\},$ $Y^1\left(r\right)=\left\{(i,u,l)\right|{\mathrm{\&gamma;}}_{i,u}^l\left(r\right)=1\},,$

$Z^0\left(r\right)=\left\{(i,j,u,l)\right|z_{i,j,u}^l\left(r\right)=0\},$ $Z^1\left(r\right)=\left\{(i,j,u,l)\right|z_{i,,j,u}^l\left(r\right)=1\}$

Wherein

Be the variable in the iteration people, by separating of from iteration, obtaining, the feasible zone of restriction upper strata primal problem.

The invention provides a kind of by stacked upper network layer being carried out traffic assignments and scheduling and lower floor being carried out the optimization transmission of the method realization of route planning to Business Stream in the stacked network.

Claims (6)

1. service resources optimized dispatching method based on stacked network, the method of traffic assignments and scheduling is carried out in employing at stacked upper network layer, with lower floor is carried out route planning and realizes optimum management method to stacked network service flow realizing the optimization transmission of Business Stream in the stacked network.Comprise the steps:

A, set up stacked upper network layer primal problem Optimization Model;

B, set up stacked network lower floor subproblem Optimization Model;

C, structure also obtain steps A and the feasible solution of the Optimization Model that B is set up;

The feasible zone of D, minimizing upper strata primal problem Optimization Model.

2. according to the method for claim 1, it is characterized in that for described steps A: set up stacked upper network layer primal problem Optimization Model, this model is used for service dispatching and transmission link distributes, and the link capacity constraints in the transmission course at one time is $\begin{array}{c}{\mathrm{\&theta;}}_{j,l}-t_{j,l}^u+S(1-z_{i,j,u}^l)+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{i,l}\\ \&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{j,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}}\end{array}$ With $\begin{array}{c}{\mathrm{\&theta;}}_{i,l}-t_{i,l}^u+{\mathrm{Sz}}_{i,j,u}^l+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{i,l},\\ \&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}}\end{array}$ , wherein With Be constant, θ _{I, l}For Business Stream i at transmitting scene l, the operation deadline,

For in the operational processes time of transmitting scene l and node u Business Stream i,

Be binary variable, if in time τ, Business Stream k is from node n _iTo n _j, then

Otherwise,

Be binary variable,, carry out the operation of Business Stream i, then if at transmitting scene

Otherwise,

Be binary variable, if at node u, transmitting scene l, Business Stream i is processed prior to Business Stream j, then

Otherwise,

Therefore upper strata primal problem Optimization Model can be described as $\begin{array}{c}\min \underset{i\&Element;J}{\mathrm{\&Sigma;}}{w}_i{T}_i\\ T_i=\max \{0,\left(c_{i,o-D_i}\right)\}\end{array},$

$s.t.z_{i,j,u}^l\&le;{\mathrm{\&gamma;}}_{i,u}^l,\&ForAll;u\&Element;M_{i,l},\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(1\right)$

$z_{i,j,u}^l\&le;{\mathrm{\&gamma;}}_{j,u}^l,\&ForAll;u\&Element;M_{j,l},\&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(2\right)$

$\underset{u\&Element;M_{i,l}}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l=1,\&ForAll;i\&Element;J,\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(3\right)$

$\begin{array}{c}{\mathrm{\&theta;}}_{j,l}-t_{j,l}^u+S(1-z_{i,j,u}^l)+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{i,l},\\ \&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{j,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},\end{array}---\left(4\right)$

$\begin{array}{c}{\mathrm{\&theta;}}_{i,l}-t_{i,l}^u+{\mathrm{Sz}}_{i,j,u}^l+S(2-{\mathrm{\&gamma;}}_{i,u}^l-{\mathrm{\&gamma;}}_{j,u}^l)\&GreaterEqual;{\mathrm{\&theta;}}_{j,l},\\ \&ForAll;i,j\&Element;J|i\&NotEqual;j,\&ForAll;u\&Element;M_{\mathrm{il},}\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}}\end{array}---\left(5\right)$

${\mathrm{\&theta;}}_{j,p\left(l\right)}-t_{j,p\left(l\right)}^u\&GreaterEqual;{\mathrm{\&theta;}}_{j,l},\&ForAll;j\&Element;J,\&ForAll;u\&Element;M_{j,l},\&ForAll;l\&Element;P\&cup;P_{\mathrm{AGV}},---\left(6\right)$

$\underset{n_j\&NotElement;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}}^k=0,\&ForAll;k\&Element;V,\&ForAll;n_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H,---\left(7\right)$

$\underset{n_j\&Element;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}}^k\&le;1,\&ForAll;k\&Element;V,{\&ForAll;n}_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H,---\left(8\right)$

$\underset{n_j\&Element;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_j,n_i,\mathrm{\&tau;}}^k=\underset{n_l\&Element;E_{n_i}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}+1,}^k\&ForAll;k\&Element;V,{\&ForAll;n}_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H-1,---\left(9\right)$

$\underset{n_j\&Element;E_{v_k}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{v_k,n_j,1}^k=1,\&ForAll;k\&Element;V,---\left(10\right)$

$\underset{k\&Element;V}{\mathrm{\&Sigma;}}\underset{n_j\&Element;E}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_j{n}_i,\mathrm{\&tau;}}^k\&le;1,\&ForAll;n_i\&Element;E,1\&le;\mathrm{\&tau;}\&le;H,---\left(11\right)$

$\underset{k\&Element;V}{\mathrm{\&Sigma;}}({\mathrm{\&eta;}}_{n_i,n_j,\mathrm{\&tau;}}^k+{\mathrm{\&eta;}}_{n_j,n_i,\mathrm{\&tau;}}^k)\&le;1,\&ForAll;n_i\&Element;E,\&ForAll;n_j\&Element;E_{n_i},1\&le;\mathrm{\&tau;}\&le;H,---\left(12\right)$

$\underset{n_j\&Element;E_{s_i^l}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{s_i^l,n_j,\mathrm{\&tau;}+1}^k\&GreaterEqual;{\mathrm{\&alpha;}}_{i,\mathrm{\&tau;},k,}^l\&ForAll;k\&Element;V,\&ForAll;l\&Element;P_{\mathrm{AGV}},1\&le;\mathrm{\&tau;}\&le;H-1,---\left(13\right)$

$\underset{n_j\&Element;E_{F_i^l}}{\mathrm{\&Sigma;}}{\mathrm{\&eta;}}_{n_j,F_i^l,\mathrm{\&tau;}}^k\&GreaterEqual;{\mathrm{\&beta;}}_{i,\mathrm{\&tau;},k}^l,\&ForAll;k\&Element;V,\&ForAll;l\&Element;P_{\mathrm{AGV}},1\&le;\mathrm{\&tau;}\&le;H,---\left(14\right)$

${\mathrm{\&theta;}}_{i,l}-t_{i,l}+1\&le;\mathrm{\&tau;}+S(1-{\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l),---\left(15\right)$

${\mathrm{\&theta;}}_{i,l}-t_{i,l}+1>\mathrm{\&tau;}-{\mathrm{S\&mu;}}_{i,\mathrm{\&tau;}}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;l\&Element;P_{\mathrm{AGV}},---\left(16\right)$

${\mathrm{\&theta;}}_{i,l}+S(1-{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l)\&GreaterEqual;\mathrm{\&tau;},---\left(17\right)$

${\mathrm{\&theta;}}_{i,l}<\mathrm{\&tau;}+S{\stackrel{-}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;l\&Element;P_{\mathrm{AGV}},---\left(18\right)$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&alpha;}}_{i,\mathrm{\&tau;},u}^l)\&GreaterEqual;2,$ (19)

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;1+{\mathrm{S\&alpha;}}_{i,t,u}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},$

${\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&beta;}}_{i,\mathrm{\&tau;},u}^l)\&GreaterEqual;2,$ (20)

${\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;1+S{\mathrm{\&beta;}}_{i,t,u}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l+S(1-{\mathrm{\&sigma;}}_{i,\mathrm{\&tau;},k}^l)\&GreaterEqual;3,---\left(21\right)$

${\mathrm{\&mu;}}_{i,\mathrm{\&tau;}}^l+{\stackrel{\&OverBar;}{\mathrm{\&mu;}}}_{i,\mathrm{\&tau;}}^l+{\mathrm{\&gamma;}}_{i,u}^l\&le;2+{\mathrm{S\&sigma;}}_{i,\mathrm{\&tau;},k,}^l,\&ForAll;i\&Element;J,1\&le;\mathrm{\&tau;}\&le;H,\&ForAll;u\&Element;M_{i,l},\&ForAll;l\&Element;P_{\mathrm{AGV}},---\left(22\right).$

3. according to the method for claim 1, it is characterized in that for described step B: set up stacked network lower floor subproblem Optimization Model, be used to carry out the selection in path, this step is satisfying under the condition of steps A, judges whether the non-conflict route of multi-path transmission exists.Exist if satisfy the route of above-mentioned condition A, then go to step D.In stacked network, do not exist if satisfy the route of above-mentioned condition A, then adopt distributed routing algorithm, obtain the optimum transmission time and postpone punishment.The zero-time and the deadline of professional transmission are set then,

And σ _{I, τ, u}, in satisfy condition (19)-(23) afterwards, the solving-optimizing model:

S.t. (7)-(14).In stacked network, increase traffic flow unit gradually, by in remaining Business Stream, inserting first traffic flow unit, produce k candidate sequence, in this sequence for current separating, select the sequential value of overall deadline local optimum, upgrade selecteed locally optimal solution and separate as current.If k=n then can obtain optimal solution, otherwise, then go to step C.

4. according to the method for claim 1, it is characterized in that:, then need the feasible solution of the Optimization Model that construction step A and B set up in this step if the Optimization Model that steps A and B are set up has infeasible solution for described step C.Substep a. finds the solution stacked network lower floor subproblem, calculate nearest professional propagation delay time, if constraints does not exist, service condition H (professional arrive time lag by the start node of separating acquisition of stacked network lower floor subproblem) then in the transmission start time of separating acquisition by stacked upper network layer primal problem, enter substep b, otherwise, constraints does not exist, then service condition I (professional arrive time lag of separating terminal point by stacked network lower floor subproblem in being separated by the optimum operation of separating acquisition of upper strata primal problem) enters substep c.Substep b. postpones to handle to Business Stream ω makes the constraints of other Business Stream be satisfied, utilization B _{ω, ω+1}=(θ _ω+1-t _ω+1)-(θ _ω+ R (F _{ω, j}, S _{ω+1, l})) calculate time of delay of next Business Stream ω+1, wherein R (F _{ω, j}, s _{ω+1, l}) be the receiving node F of Business Stream ω _{ω, l}Start node s with Business Stream ω+1 _{ω+1, l}Between minimum range, if feasible solution exists, then stop to calculate, otherwise, then be back to substep a, delay if possible is 0, then goes to substep d.Making the required minimum growth time delay of the transmission deadline of Business Stream ω is d _ω, B _{ω, ω+1}With

In minimum value.The use last time ${\mathrm{\&theta;}}_{\mathrm{\&omega;}}+\min \{d_{\mathrm{\&omega;}},\min \{d_{\mathrm{\&omega;}},\min \{({\mathrm{\&theta;}}_{\mathrm{\&omega;}}^s-t_{\mathrm{\&omega;}}^s)-{\mathrm{\&theta;}}_{\mathrm{\&omega;}},B_{\mathrm{\&omega;}+1}\}\left\}\right\}$ The more transmission deadline of new service flow ω, and use $t_{\mathrm{\&omega;}}+\min \{d_{\mathrm{\&omega;}},\min \{d_{\mathrm{\&omega;}},\min \{({\mathrm{\&theta;}}_{\mathrm{\&omega;}}^s-t_{\mathrm{\&omega;}}^s)-{\mathrm{\&theta;}}_{\mathrm{\&omega;}},B_{\mathrm{\&omega;}+1}\}\left\}\right\}$ More the transmission time of new service flow ω, find the solution for lower floor's subproblem Optimization Model, if feasible solution exists, then algorithm stops, otherwise is back to substep a, and be 0 time of delay if possible, then goes to substep d.Substep c, more the transmission deadline of new service flow ω.Substep d, the orlop subproblem is found the solution, and d time of delay of computing service stream ω _ωSubstep e uses the transmission deadline θ of last Business Stream ω _ω+ d _ωAs its transmission deadline θ next time _ωUse this θ _α+ d _αAs business transmission deadline θ next time _α, wherein $d_{\mathrm{\&alpha;}}=\left\{\begin{array}{cc}{d}_{\mathrm{\&alpha;}-1}& (B_{\mathrm{\&alpha;}-1,\mathrm{\&alpha;}}=0)\\ \max \left(d_{\mathrm{\&alpha;}-1}{-B}_{\mathrm{\&alpha;}-\mathrm{1,}\mathrm{\&alpha;},0}\right)& (B_{\mathrm{\&alpha;}-1,\mathrm{\&alpha;}}\&NotEqual;0)\end{array}\right.,$ $d_{{\mathrm{\&alpha;}}^{\&prime;}}^s=\max \{0,d_e-\min \{({\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}^s-t_{{\mathrm{\&alpha;}}^{\&prime;}}^s)+d_e-{\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}-1,}^s({\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}^s-t_{{\mathrm{\&alpha;}}^{\&prime;}}^s)+d_e-{\mathrm{\&theta;}}_{{\mathrm{\&alpha;}}^{\&prime;}}\left\}\right\}$ For satisfying the minimum delay of sequence constraints.Lower floor's subproblem Optimization Model is found the solution,, then stopped if can obtaining feasible solution, otherwise, go to steps A.Substep f uses last deadline and transmission time θ _ω+ d _ωAnd t _ω+ d _ωAs deadline and the transmission time of Business Stream e next time, the lower floor subproblem is found the solution, if feasible solution exists, then stop, otherwise, then be back to substep a.

5. according to the method for claim 1, it is characterized in that for described step C: convergence domain is assessed, and the difference of obtaining the upper limit and the lower limit that obtains from steps A from step B and C is ε, and enters step D.

6. according to the method for claim 1, it is characterized in that for described step D: get rid of previously obtd possible arrangement by using the integer cutting method, this possible arrangement can reduce upper strata primal problem Optimization Model feasible zone.

Be separating of the r time iteration acquisition.For making primal problem obtain to dispose at the difference of variable, use the integer cutting method in the upper strata subproblem,

$\underset{(i,u,l)\&Element;Y^1\left(r\right)}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l-\underset{(i,u,l)\&Element;Y^0\left(r\right)}{\mathrm{\&Sigma;}}{\mathrm{\&gamma;}}_{i,u}^l\&le;\left|Y^1\left(r\right)\right|,\&ForAll;r,$

$\underset{(i,j,u,l)\&Element;z^1\left(r\right)}{\mathrm{\&Sigma;}}{z}_{i,j,u}^l-\underset{(i,j,u,l)\&Element;z^0\left(r\right)}{\mathrm{\&Sigma;}}{z}_{i,j,u}^l\&le;\left|z^1\left(r\right)\right|,\&ForAll;r,$

$Y^0\left(r\right)=\left\{(i,u,l)\right|{\mathrm{\&gamma;}}_{i,u}^l\left(r\right)=0\},$ $Y^1\left(r\right)=\left\{(i,u,l)\right|{\mathrm{\&gamma;}}_{i,u}^l\left(r\right)=1\},,$

$Z^0\left(r\right)=\left\{(i,j,u,l)\right|z_{i,j,u}^l\left(r\right)=0\},$ $Z^1\left(r\right)=\left\{(i,j,u,l)\right|z_{i,,j,u}^l\left(r\right)=1\}$

Wherein

Be the variable in the iteration people, by separating of from iteration, obtaining, the feasible zone of restriction upper strata primal problem.

Images