CN108234617A - A kind of resource dynamic dispatching method under the mixing cloud mode towards electric system - Google Patents
A kind of resource dynamic dispatching method under the mixing cloud mode towards electric system Download PDFInfo
- Publication number
- CN108234617A CN108234617A CN201711429164.6A CN201711429164A CN108234617A CN 108234617 A CN108234617 A CN 108234617A CN 201711429164 A CN201711429164 A CN 201711429164A CN 108234617 A CN108234617 A CN 108234617A
- Authority
- CN
- China
- Prior art keywords
- node
- task
- concentration
- pheromone concentration
- ant
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1023—Server selection for load balancing based on a hash applied to IP addresses or costs
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/56—Provisioning of proxy services
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system, steps：Pheromone concentration is initialized；Operation is taken out in job queue, x ant of identical quantity is sent to each task, and starts timer, every ant random selection, according to the movement rule of ant, finds enabled node from some node in private clound；Judge that enabled node is concentrated with the presence or absence of node, to be able to carry out the task if there are enabled nodes in private clound in the presence of if, assign the task to enabled node and concentrate the minimum node of pheromone concentration, reduce the pheromone concentration of the node；Private clound does not perform the task if available section point set is sky, applies for that resource is performed into public cloud, and the public cloud virtual machine node applied is added to enabled node concentrates；After the completion of task, restore the pheromone concentration of the node；Whether task smoothly completes, if completing, then increases the pheromone concentration of the node, if failure, reduces the concentration of pheromones.
Description
Technical field
The present invention relates to a kind of power system resource dispatching method, especially with regard to one kind towards electric system mixed cloud mould
Resource dynamic dispatching method under formula.
Background technology
At present, intelligent grid is being built under the new situation, the opportunities and challenges that power informatization is faced with are unprecedented
's.The huge supersystem that electric system has distributed constant as one, it does not require nothing more than reliability height, while should also have
There is the characteristic of the characteristics of real-time and NATURAL DISTRIBUTION.Also, production to electric power and management also should comply with " differentiated control,
The requirement of hierarchical control and distribution process ".
With flourishing for intelligent grid, it is various each that nowadays existing electric system cannot increasingly meet user already
The demand of sample has also been proposed many requirements such as high-quality, efficient, green, interactive, self-healing property, information sharing on this basis.
Therefore the scheduling that is rationalized to the resource in electric system cloud platform, ensure cloud platform can Effec-tive Function while,
The utilization rate of resource can also be improved, the requirement of power grid could be met in this way, be push electric system move towards intelligence, greenization and
The important channel of sharedization.
And scheduling of resource is to improve the performance of electric system cloud platform and realize that the load of electric system cloud computing is equal
One of important mechanisms of weighing apparatus.According to the deployment mode of cloud computing, cloud can be divided into private clound, public cloud, community cloud and mixing
Four type of cloud.So-called mixed cloud exactly to a certain extent combines private clound and public cloud.Therefore mixed cloud is not only
The scheduling method of private clound offer can be provided, flexibly can also expand oneself using the service of public cloud offer
Service, so as to effectively improve electric system reply access fluctuation etc. emergency cases ability.In addition to this, it mixes
It closes cloud and can also be used to effectively cope with proactive explosion type and access, the electricity consumption just such as during summer hot weather is high
Peak period or large-scale activity need to use a large amount of electrical equipment etc. during carrying out, and are likely under this situation to entire
Electric system brings very big impact.At this point, on the basis of private clound is used, the service providers of public cloud can also be used
The free service that is there is provided reaches cooperation agreement with the service providers of public cloud, provides service together, exploitation is based on mixing
The operation mode of cloud.When electricity usage amount is smaller, using only the service of private clound, and when usage amount is larger, can incite somebody to action
The problem of most usage amount is transferred to the service of public cloud up, and such explosion type accesses just is readily solved.
Currently for the research of the resource scheduling algorithm of cloud computing, the scheduling of resource mostly in single cloud is calculated
Method, it is then fewer and fewer to the research of the resource scheduling algorithm of mixed cloud.According to the actual demand of user, it is necessary to provide a kind of needle
To the resource regulating method of mixed cloud.
Invention content
In view of the above-mentioned problems, the object of the present invention is to provide the resource dynamics under a kind of mixing cloud mode towards electric system
Dispatching method, this method can make privately owned cloud resource maximize the use, effectively reduce entreprise cost, reduce the wave of resource
Take.
To achieve the above object, the present invention takes following technical scheme：Under a kind of mixing cloud mode towards electric system
Resource dynamic dispatching method, it is characterised in that include the following steps：1) user submits operation to define each void to proxy server
The pheromones of plan machine node, initialize the pheromone concentration of each virtual machine node；2) proxy server successively from
Operation is taken out in the job queue that family is submitted, the task quantity that is included according to the operation sends ant, and each task is sent
Go out x ant of identical quantity, and start timer, every ant random selection is from some node, according to ant
Movement rule finds enabled node in private clound, if finding, is placed in enabled node concentration, while find available section
The ant of point returns to scheduler, is then continually looked for if not finding；3) when timer arrival presets the time, judge available
With the presence or absence of node in set of node, if there are node, show that there are enabled nodes in private clound to be able to carry out the task, so
Enabled node is assigned the task to afterwards concentrates the minimum node of pheromone concentration, and reduce the pheromone concentration of the node；It if can
It is empty with set of node, then shows that private clound is not enough to perform the task, need to apply for resource into public cloud at this time, performs this
Business, and the public cloud virtual machine node applied is added to enabled node and is concentrated；4) after tasks carrying is completed, release accounts for
Resource then restores the pheromone concentration of the node；Judge whether task smoothly completes again, if smoothly completing, just increase again
The pheromone concentration of the node if tasks carrying fails, just reduces the concentration of pheromones；5) whether judge the task of the work in the industry
Completion all is performed, is not completed, next task is searched and performs node, and return to step 2)；Continue to judge to make if completing
Whether operation all performs completion in industry queue, is, terminates, on the contrary then return to step 2), step 2) is repeated to step
4), until task all performs completion.
Further, in the step 1), the resource in each virtual machine node has CPU, memory, hard disk and network bandwidth,
It is VM the resource definition of virtual machinei={ bi,ui,mi,hi,di, i is node number, i>0；B represents the bandwidth of network；U is represented
The quantity of CPU；M represents the capacity of memory；H represents the capacity of hard disk；D represents the processing capacity of CPU, unit MPIS；For
Each parameter all sets a critical value, and if it exceeds the critical value, then unify to carry out using critical value as value
It calculates, wherein critical value is set separately as follows：
bmax=b0,umax=u0,mmax=m0,hmax=h0,dmax=d0；
In formula, b0Critical value for network bandwidth；u0Critical value for CPU quantity；m0Critical value for memory size；h0
Critical value for hard-disk capacity；d0Critical value for CPU processing capacities；
Then, the pheromone concentration C of CPU computing capabilitysic(0) it is：
The pheromone concentration C of memory sizeim(0) it is：
The pheromone concentration C of hard-disk capacityih(0) it is expressed as：
The pheromone concentration C of network bandwidthib(0) it is expressed as：
The pheromone concentration of node i is the weighted sum of the pheromone concentration of parameters：
Ci=a*Cic+b*Cim+c*Cih+d*Cib,
Wherein, a+b+c+d=1.
Further, in the step 2), the movement rule of ant is：Information interchange is carried out by pheromones between ant,
What ant progress information interchange relied on is the size of pheromone concentration on each node, dense in the pheromones for representing virtual machine node
In the matrix Pheromone of degree, therefore to consider the estimation of the concentration and new task of pheromones on this node and perform the time,
So as to find the node based on maximum probability in the node to the node accessed next time.
Further, the probability is：
In formula, PijFor the probability of node j that should be reached in node i selection next time；CjIt is to detect node in node i
The concentration of pheromones on j；PTj=PTj(Task (t)) i.e. new task Task is performed in estimation of the t moment on node j
Time；NiIt is the set for the path node that ant can reach；E is the abutment points of node i；α is important to represent pheromone concentration
The parameter of degree；β is to represent that estimation performs the parameter of time significance level；AllowediRepresent that node i allows the node accessed
Set.
Further, estimation execution time model is established, task performs time, estimation in the estimation of t moment on calculate node j
Execution time model is：
Wherein, RTnIt is the actual execution time of n-th of task；PTnIt is the estimation execution time of n-th of task；PTn+1It is
The estimation of (n+1)th task, that is, current task performs the time；It is RTnWeights, then：
Wherein, PT0Estimation for initiating task performs the time.
Further, enabled node is found, and adds it to enabled node concentration, it is assumed that ant is in node i, and through sentencing
Disconnected node i is not enabled node, and process is as follows：2.1) node is deposited into forbidden list, it is not available to represent the node
Node prevents other ants from accessing again；2.2) it is known to have node j, node k and node l, root with the adjacent node of node i
According to the movement rule of ant, ant is calculated respectively from node i next time to the probability of the node accessed being adjacent；2.3) it selects
Go out the node accessed next time that the node of maximum probability is selected as ant, it is assumed that the maximum probability of node j, so the ant
The node accessed next time be node j；2.4) whether decision node j is enabled node, if enabled node, then by the node
It is added to enabled node concentration；If 2.5) node j is not available node, which repeats step 2.1) to step 2.4),
It continually looks for, until timer arrival presets the time.
Further, in the step 3), judge whether enabled node concentration has enabled node, and change in tasks carrying
The concentration of nodal information element, process are as follows：If there is enabled node, show that the task can be performed in private clound, therefore not
The resource of application public cloud must be gone, is concentrated in enabled node, pheromone concentration is selected in the set most according to optimal adaptation algorithm
Small node is most available node, distributes to tasks carrying；The pheromone concentration of the node is reduced at this time, reduce the node
The formula of pheromone concentration be：
Wherein, p is the volatility coefficient of pheromones, and value range isΔδtIt represents occupied in the task of execution
The ratio magnitude of resource；Ci preRepresent the pheromone concentration of the node when being also not carried out the task；Ci onIt represents performing this
The pheromone concentration of the node in business；
After the completion of tasks carrying, the pheromone concentration of the node then returns to original concentration：
In formula,Represent that tasks carrying completes the pheromone concentration of posterior nodal point.
Further, in the step 4), when Mission Success, which performs, to be completed, increase the concentration of the node：
Ci(t2)=(1+ λ1)Ci(t1), λ1>0
Wherein, Ci(t2) represent tasks carrying after the completion of, the pheromone concentration of the node；Ci(t1) represent tasks carrying before
The pheromone concentration of the node；λ1The increased amplitude of nodal information element, the weight depending on task after expression Mission Success performs
Want degree.
Further, in the step 4), completion is smoothly performed when task is no, reduces the concentration of the node：
Ci(t2)=(1- λ2)Ci(t1), λ2>0
Wherein, Ci(t2) represent tasks carrying after the completion of, the pheromone concentration of the node；Ci(t1) represent tasks carrying before
The pheromone concentration of the node；λ2The amplitude that the nodal information element reduces after expression tasks carrying failure, the weight depending on task
Want degree.
The present invention has the following advantages due to taking above technical scheme：1st, the present invention is newest according to cloud computing technology
State of development and the business characteristic of electric system and the demand for development of intelligent grid, based on ant group algorithm, to virtual
The pheromone concentration of machine node is compared, and enabled node is found in the time is preset, and then judges whether task can be with
It is performed in private clound, the task in user job is made to be carried out in existing private clound as far as possible, if private clound cannot perform
During the user job, application public cloud is just gone, so as to reduce entreprise cost, reduces the waste of resource.2nd, the present invention can in searching
During with node, the pheromone concentration size and task that consider on private clound node perform time, system in the estimation of the node
Determine ant movement rule.It finds out all enabled nodes in the time is preset and is added to enabled node concentration, further according to
Optimal adaptation algorithm is focused to find out node of the node of concentration minimum as distribution task from enabled node, if not finding available
Node then goes application public cloud, privately owned cloud resource can be made to maximize the use.
Description of the drawings
Fig. 1 is the overall flow schematic diagram of the present invention；
Fig. 2 is that enabled node finds path profile in the present invention.
Specific embodiment
Under normal conditions, can all there be the private clound of oneself inside power grid enterprises, the use cost of private clound is fixed.When
User's increased situation suddenly using the cloud service is had, when being more than the load limit that can be undertaken of private clound,
It can go resource of the application using public cloud.Cai Nengshi enterprises spend minimum cost in this way, to meet the mission requirements of itself.Cause
This, it is main problem to be solved by this invention that the privately owned cloud resource that enterprise possesses how to be made, which to maximize the use,.Below
The present invention is described in detail in conjunction with the accompanying drawings and embodiments.
As shown in Figure 1, the present invention provides the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system, it should
Method is compared the concentration of pheromones based on ant group algorithm, to private clound within the timer preset time
The lookup of enabled node is carried out, and then judges whether some task can perform in private clound, it is desired nonetheless to go to apply publicly-owned
A kind of resource scheduling algorithm of cloud resource.In the assignment procedure, consider private clound enabled node on pheromone concentration size and
Task the node estimation perform the time, so as to find out all enabled nodes in the time is preset and them all
It is added in available section point set.Then it is dense pheromones to be focused to find out from enabled node further according to optimal adaptation algorithm (Best Fit)
Spend node of the minimum node as distribution task.The present invention specifically includes following steps：
1) user submits operation the pheromones of each virtual machine node to be defined, to each virtual machine section to proxy server
The pheromone concentration of point is initialized；
The quantity for being assumed to be the ant that each task is sent is x, by the pheromone concentration square of all virtual machine nodes
Battle array Pheromone is represented, then sets a forbidden list (Useless) again to store the node that ant had accessed, and
The node inside this table is not visited again in later search.Setting one again allows the node table (Allowed) accessed to deposit
Store up the node that can also be accessed.
Resource in each virtual machine node has CPU, memory, hard disk and network bandwidth, with the capacity of water of these resources
As the determinant initialized to nodal information element concentration.Therefore it is VM the resource definition of virtual machinei, and retouched
State for：VMi={ bi,ui,mi,hi,di, i is node number, i>0.Wherein b represents the bandwidth of network；U represents the quantity of CPU；m
Represent the capacity of memory；H represents the capacity of hard disk；D represents the processing capacity of CPU, unit MPIS.For each
Parameter all sets a critical value (being exactly that the maximum value of parameter inside all resources that can be utilized is set as critical value),
And if it exceeds the critical value, then unify to be calculated using critical value as value, wherein critical value is set separately as follows：
bmax=b0,umax=u0,mmax=m0,hmax=h0,dmax=d0 (1)
In formula, b0Critical value for network bandwidth；u0Critical value for CPU quantity；m0Critical value for memory size；h0
Critical value for hard-disk capacity；d0Critical value for CPU processing capacities；
Then, the pheromone concentration C of CPU computing capabilitysic(0) it is：
The pheromone concentration C of memory sizeim(0) it is：
The pheromone concentration C of hard-disk capacityih(0) it is expressed as：
The pheromone concentration C of network bandwidthib(0) it is expressed as：
The pheromone concentration of node i is the weighted sum C of the pheromone concentration of parametersi：
Ci=a*Cic+b*Cim+c*Cih+d*Cib, (6)
Wherein, a+b+c+d=1.
2) proxy server takes out operation in the job queue that user submits successively, being included according to the operation for task
Quantity sends ant, x ant of identical quantity is sent to each task, and start timer, every ant random selection
From some node, according to the movement rule of ant, enabled node is found in private clound, if finding, is placed in
Enabled node is concentrated, while the ant for finding enabled node returns to scheduler, is then continually looked for if not finding；
Wherein, the movement rule of ant is as follows：
Information interchange can be carried out by pheromones between ant, so as to find shortest road by cooperating with each other
Diameter；Due to ant carry out that information interchange relies on be pheromone concentration on each node size, representing virtual machine node
In the matrix Pheromone of pheromone concentration, therefore to consider the estimation of the concentration and new task of pheromones on this node
The time is performed, so as to find the node based on maximum probability in the node to the node accessed next time.The probability is：
In formula, PijFor the probability of node j that should be reached in node i selection next time；CjIt is to detect node in node i
The concentration of pheromones on j；PTj=PTj(Task (t)) i.e. new task Task is performed in estimation of the t moment on node j
Time；NiIt is the set for the path node that ant can reach；E is the abutment points of node i；PTnIt is that the estimation of n-th of task is held
The row time；α is the parameter for representing pheromone concentration significance level；β is to represent that estimation performs the parameter of time significance level,
AllowediRepresent that node i allows the set of the node accessed.
In the present embodiment, estimation execution time model is established, task is when the estimation of t moment performs on calculate node j
Between：
Estimate that execution time model is：
Wherein, RTnIt is the actual execution time of n-th of task；PTnIt is the estimation execution time of n-th of task；PTn+1It is
The estimation of (n+1)th task, that is, current task performs the time；It is RTnWeights, be by formula (8) is deployable：
Wherein, PT0Estimation for initiating task performs the time.
Due toTherefore in time, the execution of the historic task nearer apart from current task
Perform that time time effects are bigger the time to the estimation of current task, and time task more remote holds the estimation of current task
The influence of row time is with regard to smaller, so PT0RT can be used0To replace.The practical rule of system is met using above method, is had
Conducive to the accuracy of estimation.
3) when timer arrival presets the time, judge that enabled node is concentrated with the presence or absence of node, if there are node,
Then show that there are enabled nodes in private clound to be able to carry out the task, then assign the task to enabled node and concentrate pheromones
The minimum node of concentration, and reduce the pheromone concentration of the node.If available section point set is sky, show that private clound is not enough to hold
The row task, needs to apply for resource into public cloud at this time, performs the task, and the public cloud virtual machine node applied is added
Enter to enabled node and concentrate；
Wherein, judge whether enabled node concentration has enabled node, and the dense of nodal information element is changed in tasks carrying
Degree, process are as follows：
If there is enabled node, show that the task can be performed in private clound, therefore the money of application public cloud need not be gone
Source is concentrated in enabled node, and the node of pheromone concentration minimum in the set is selected according to optimal adaptation algorithm (Best Fit)
For most available node, tasks carrying is distributed to.Because the node can occupy certain resource in the task of execution, lead to the section
The ability that point performs other tasks reduces, and to reduce the pheromone concentration of the node at this time, prevent other tasks still according to original
The pheromone concentration come has carried out the node judgement of mistake and the node has been selected to perform task, it is possible to cause to rush
It is prominent.The formula for reducing the pheromone concentration of the node is：
Wherein, p is the volatility coefficient of pheromones, and value range isΔδtIt represents occupied in the task of execution
The ratio magnitude of resource；Ci preRepresent the pheromone concentration of the node when being also not carried out the task；Ci onIt represents performing this
The pheromone concentration of the node in business.
After the completion of tasks carrying, the pheromone concentration of the node then returns to original concentration：
In formula,Represent that tasks carrying completes the pheromone concentration of posterior nodal point.
4) resource is occupied during tasks carrying, the nodal information element concentration should be reduced at this time.After tasks carrying is completed,
The resource of release busy then restores the pheromone concentration of the node.Then judge whether task smoothly completes again, if smoothly complete
Into just increasing the pheromone concentration of the node again, ensure to attract more ants.If tasks carrying fails, pheromones are just reduced
Concentration reduces temporal waste with this；
Wherein, after tasks carrying is complete, the concentration of the pheromones of the node is changed, process is as follows：
4.1) when Mission Success, which performs, to be completed, increase the concentration of the node：
Ci(t2)=(1+ λ1)Ci(t1), λ1>0 (12)
Wherein, Ci(t2) represent tasks carrying after the completion of, the pheromone concentration of the node；Ci(t1) represent tasks carrying before
The pheromone concentration of the node；λ1The increased amplitude of nodal information element, the weight depending on task after expression Mission Success performs
Want degree.
4.2) completion is smoothly performed when task is no, just reduces the concentration of the node：
Ci(t2)=(1- λ2)Ci(t1), λ2>0 (13)
Wherein, λ2The amplitude that the nodal information element reduces after expression tasks carrying failure, the significance level depending on task.
5) judge whether the task of the work in the industry all performs completion, do not complete, search next task and perform section
Point, and return to step 2)；Continue to judge whether operation all performs completion in job queue, is to terminate if completing, it is on the contrary
Then return to step 2), step 2) is repeated to step 4), until task all performs completion.
Above-mentioned steps 2) in, as shown in Fig. 2, finding enabled node, and add it to enabled node concentration, it is assumed that ant
It is not enabled node in node i, and through decision node i, process is as follows：
2.1) node is deposited into forbidden list (Useless), it is not enabled node to represent the node, prevents other ants
It accesses again；
2.2) it is known to have node j, node k and node l etc. with the adjacent node of node i, according to the mobile rule of ant
Then, ant is calculated respectively from node i next time to the probability of the node accessed being adjacent；
2.3) node accessed next time that the node of maximum probability is selected as ant is selected, it is assumed that the probability of node j
Maximum, so the node accessed of the ant is node j next time；
2.4) whether decision node j is enabled node, if enabled node, then the node is added to available section point set
In；
If 2.5) node j is not available node, which repeats step 2.1) to step 2.4), continually looks for, directly
Until timer arrival presets the time.
The various embodiments described above are merely to illustrate the present invention, and structure and size, installation position and the shape of each component are all can be with
It is varied from, on the basis of technical solution of the present invention, all improvement carried out according to the principle of the invention to individual part and waits
With transformation, should not exclude except protection scope of the present invention.
Claims (9)
1. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system, it is characterised in that include the following steps：
1) user submits operation the pheromones of each virtual machine node to be defined, to each virtual machine node to proxy server
Pheromone concentration is initialized；
2) proxy server takes out operation in the job queue that user submits successively, the task quantity included according to the operation
Send ant, x ant of identical quantity sent to each task, and start timer, every ant random selection from certain
One node sets out, and according to the movement rule of ant, enabled node is found in private clound, if finding, is placed in available
In set of node, while the ant for finding enabled node returns to scheduler, is then continually looked for if not finding；
3) when timer arrival presets the time, judge that enabled node is concentrated with the presence or absence of node, if there are node, table
Bright there are enabled nodes in private clound to be able to carry out the task, then assigns the task to enabled node and concentrates pheromone concentration
Minimum node, and reduce the pheromone concentration of the node；If available section point set is sky, shows that private clound is not enough to perform and be somebody's turn to do
Task needs to apply for resource into public cloud at this time, performs the task, and the public cloud virtual machine node applied is added to
Enabled node is concentrated；
4) after tasks carrying is completed, the resource of release busy then restores the pheromone concentration of the node；Judging task again is
It is no to smoothly complete, if smoothly completing, just increase the pheromone concentration of the node again, if tasks carrying fails, just reduce pheromones
Concentration；
5) judge whether the task of the work in the industry all performs completion, do not complete, search next task and perform node, and
Return to step 2)；Continue to judge whether operation all performs completion in job queue, is to terminate if completing, it is on the contrary then return
Step 2) repeats step 2) to step 4), until task all performs completion.
2. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as described in claim 1, feature
It is：In the step 1), the resource in each virtual machine node has CPU, memory, hard disk and network bandwidth, virtual machine
Resource definition is VMi={ bi,ui,mi,hi,di, i is node number, i>0；B represents the bandwidth of network；U represents the quantity of CPU；
M represents the capacity of memory；H represents the capacity of hard disk；D represents the processing capacity of CPU, unit MPIS；For each of which
A parameter all sets a critical value, and if it exceeds the critical value, then unify to be calculated using critical value as value, wherein facing
Dividing value is set separately as follows：
bmax=b0,umax=u0,mmax=m0,hmax=h0,dmax=d0；
In formula, b0Critical value for network bandwidth；u0Critical value for CPU quantity；m0Critical value for memory size；h0For hard disk
The critical value of capacity；d0Critical value for CPU processing capacities；
Then, the pheromone concentration C of CPU computing capabilitysic (0)For：
The pheromone concentration C of memory sizeim (0)For：
The pheromone concentration C of hard-disk capacityih (0)It is expressed as：
The pheromone concentration C of network bandwidthib (0)It is expressed as：
The pheromone concentration of node i is the weighted sum of the pheromone concentration of parameters：
Ci=a*Cic+b*Cim+c*Cih+d*Cib,
Wherein, a+b+c+d=1.
3. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as described in claim 1, feature
It is：In the step 2), the movement rule of ant is：Information interchange is carried out by pheromones between ant, ant carries out letter
What breath exchange relied on is the size of pheromone concentration on each node, in the matrix for the pheromone concentration for representing virtual machine node
In Pheromone, therefore to consider the estimation of the concentration and new task of pheromones on this node and perform the time, so as to find
Node based on maximum probability in the node to the node accessed next time.
4. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as claimed in claim 3, feature
It is：The probability is：
In formula, PijFor the probability of node j that should be reached in node i selection next time；CjIt is to be detected on node j in node i
Pheromones concentration；PTj=PTj(Task (t)) i.e. new task Task is when estimation of the t moment on node j performs
Between；NiIt is the set for the path node that ant can reach；E is the abutment points of node i；α is represents the important journey of pheromone concentration
The parameter of degree；β is to represent that estimation performs the parameter of time significance level；AllowediRepresent that node i allows the node accessed
Set.
5. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as claimed in claim 3, feature
It is：Estimation execution time model is established, task performs the time in the estimation of t moment on calculate node j, and estimation performs time mould
Type is：
Wherein, RTnIt is the actual execution time of n-th of task；PTnIt is the estimation execution time of n-th of task；PTn+1It is (n+1)th
The estimation of a task, that is, current task performs the time；It is RTnWeights, then：
Wherein, PT0Estimation for initiating task performs the time.
6. a kind of resource dynamic dispatching side under mixing cloud mode towards electric system as described in claim 1 or 3 or 4 or 5
Method, which is characterized in that find enabled node, and add it to enabled node concentration, it is assumed that ant is in node i, and through sentencing
Disconnected node i is not enabled node, and process is as follows：
2.1) node is deposited into forbidden list, it is not enabled node to represent the node, prevents other ants from accessing again；
2.2) it is known to have node j, node k and node l with the adjacent node of node i, according to the movement rule of ant, count respectively
Ant is calculated from node i next time to the probability of the node accessed being adjacent；
2.3) node accessed next time that the node of maximum probability is selected as ant is selected, it is assumed that the maximum probability of node j,
So the node accessed next time of the ant is node j；
2.4) whether decision node j is enabled node, if enabled node, then the node is added to enabled node and concentrated；
If 2.5) node j is not available node, which repeats step 2.1) to step 2.4), continually looks for, Zhi Daoji
When device arrival preset the time until.
7. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as described in claim 1, feature
It is：In the step 3), judge whether enabled node concentration has enabled node, and nodal information element is changed in tasks carrying
Concentration, process is as follows：
If there is enabled node, show that the task can be performed in private clound, therefore the resource of application public cloud need not be gone, can
With in set of node, the node that pheromone concentration minimum in the set is selected according to optimal adaptation algorithm is most available node, point
Dispensing tasks carrying；The pheromone concentration of the node is reduced at this time, and the formula for reducing the pheromone concentration of the node is：
Wherein, p is the volatility coefficient of pheromones, and value range isΔδtRepresent the occupied resource in the task of execution
Ratio magnitude；Ci preRepresent the pheromone concentration of the node when being also not carried out the task；Ci onIt represents in the task is performed
The pheromone concentration of the node；
After the completion of tasks carrying, the pheromone concentration of the node then returns to original concentration：
In formula,Represent that tasks carrying completes the pheromone concentration of posterior nodal point.
8. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as described in claim 1, feature
It is：In the step 4), when Mission Success, which performs, to be completed, increase the concentration of the node：
Ci(t2)=(1+ λ1)Ci(t1), λ1>0
Wherein, Ci(t2) represent tasks carrying after the completion of, the pheromone concentration of the node；Ci(t1) represent the node before tasks carrying
Pheromone concentration；λ1The increased amplitude of nodal information element, the significance level depending on task after expression Mission Success performs.
9. the resource dynamic dispatching method under a kind of mixing cloud mode towards electric system as described in claim 1, feature
It is：In the step 4), completion is smoothly performed when task is no, reduces the concentration of the node：
Ci(t2)=(1- λ2)Ci(t1), λ2>0
Wherein, Ci(t2) represent tasks carrying after the completion of, the pheromone concentration of the node；Ci(t1) represent the node before tasks carrying
Pheromone concentration；λ2The amplitude that the nodal information element reduces after expression tasks carrying failure, the significance level depending on task.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711429164.6A CN108234617B (en) | 2017-12-26 | 2017-12-26 | Dynamic resource scheduling method in hybrid cloud mode for power system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711429164.6A CN108234617B (en) | 2017-12-26 | 2017-12-26 | Dynamic resource scheduling method in hybrid cloud mode for power system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108234617A true CN108234617A (en) | 2018-06-29 |
| CN108234617B CN108234617B (en) | 2021-01-15 |
Family
ID=62648754
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711429164.6A Active CN108234617B (en) | 2017-12-26 | 2017-12-26 | Dynamic resource scheduling method in hybrid cloud mode for power system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108234617B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110245808A (en) * | 2019-06-25 | 2019-09-17 | 湘潭大学 | A kind of ladle furnace optimization dispatching method based on demand control |
| CN112084035A (en) * | 2020-09-21 | 2020-12-15 | 西安理工大学 | Task scheduling method and system based on ant colony algorithm |
| CN112468558A (en) * | 2020-11-16 | 2021-03-09 | 中科三清科技有限公司 | Request forwarding method, device, terminal and storage medium based on hybrid cloud |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103345657A (en) * | 2013-04-02 | 2013-10-09 | 江苏大学 | Task scheduling method based on heredity and ant colony in cloud computing environment |
| CN103731372A (en) * | 2013-12-24 | 2014-04-16 | 东南大学 | Resource supply method for service supplier under hybrid cloud environment |
| US20150223441A1 (en) * | 2014-02-13 | 2015-08-13 | The Regents Of The University Of California | Development of pheromone-assisted techniques (pat) to improve efficacy of insecticide baits targeting urban pest ant species |
-
2017
- 2017-12-26 CN CN201711429164.6A patent/CN108234617B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103345657A (en) * | 2013-04-02 | 2013-10-09 | 江苏大学 | Task scheduling method based on heredity and ant colony in cloud computing environment |
| CN103731372A (en) * | 2013-12-24 | 2014-04-16 | 东南大学 | Resource supply method for service supplier under hybrid cloud environment |
| US20150223441A1 (en) * | 2014-02-13 | 2015-08-13 | The Regents Of The University Of California | Development of pheromone-assisted techniques (pat) to improve efficacy of insecticide baits targeting urban pest ant species |
Non-Patent Citations (1)
| Title |
|---|
| 刘永: "云计算环境下虚拟机资源调度策略研究", 《山东师范大学硕士学位论文》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110245808A (en) * | 2019-06-25 | 2019-09-17 | 湘潭大学 | A kind of ladle furnace optimization dispatching method based on demand control |
| CN110245808B (en) * | 2019-06-25 | 2023-05-12 | 湘潭大学 | Ladle furnace optimal scheduling method based on demand control |
| CN112084035A (en) * | 2020-09-21 | 2020-12-15 | 西安理工大学 | Task scheduling method and system based on ant colony algorithm |
| CN112084035B (en) * | 2020-09-21 | 2024-04-30 | 西安理工大学 | Task scheduling method and system based on ant colony algorithm |
| CN112468558A (en) * | 2020-11-16 | 2021-03-09 | 中科三清科技有限公司 | Request forwarding method, device, terminal and storage medium based on hybrid cloud |
| CN112468558B (en) * | 2020-11-16 | 2021-08-20 | 中科三清科技有限公司 | Request forwarding method, device, terminal and storage medium based on hybrid cloud |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108234617B (en) | 2021-01-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110231976B (en) | Load prediction-based edge computing platform container deployment method and system | |
| CN108182105B (en) | Local dynamic migration method and control system based on Docker container technology | |
| Wei-guo et al. | Research on kubernetes' resource scheduling scheme | |
| Wided et al. | Load balancing with Job Migration Algorithm for improving performance on grid computing: Experimental Results | |
| CN105245617A (en) | Container-based server resource supply method | |
| CN103152393A (en) | Charging method and charging system for cloud computing | |
| CN104503832B (en) | A kind of scheduling virtual machine system and method for fair and efficiency balance | |
| CN106790636A (en) | A kind of equally loaded system and method for cloud computing server cluster | |
| CN104050042A (en) | Resource allocation method and resource allocation device for ETL (Extraction-Transformation-Loading) jobs | |
| CN103491024A (en) | Job scheduling method and device for streaming data | |
| CN108234617A (en) | A kind of resource dynamic dispatching method under the mixing cloud mode towards electric system | |
| CN104679594A (en) | Middleware distributed calculating method | |
| Supreeth et al. | Hybrid genetic algorithm and modified-particle swarm optimization algorithm (GA-MPSO) for predicting scheduling virtual machines in educational cloud platforms | |
| Supreeth et al. | Comparative approach for VM scheduling using modified particle swarm optimization and genetic algorithm in cloud computing | |
| CN114938372B (en) | Federal learning-based micro-grid group request dynamic migration scheduling method and device | |
| CN109343945A (en) | A kind of multitask dynamic allocation method based on contract net algorithm | |
| Zubair et al. | Integration of cloud-fog based platform for load balancing using hybrid genetic algorithm using bin packing technique | |
| CN109740870A (en) | The resource dynamic dispatching method that Web is applied under cloud computing environment | |
| CN108845886A (en) | Cloud computing energy consumption optimization method and system based on phase space | |
| Hao et al. | Evaluation of nine heuristic algorithms with data‐intensive jobs and computing‐intensive jobs in a dynamic environment | |
| CN104683480A (en) | Distribution type calculation method based on applications | |
| Yang et al. | Multi-resource allocation for virtual machine placement in video surveillance cloud | |
| CN109408230A (en) | Docker container dispositions method and system based on energy optimization | |
| CN112968962B (en) | Cloud platform storage resource scheduling method based on distributed computer cluster architecture | |
| Somasundaram et al. | Task resource allocation in grid using swift scheduler |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |