CN101169738A - Self-adaptive media content creative objective assignment method - Google Patents
Self-adaptive media content creative objective assignment method Download PDFInfo
- Publication number
- CN101169738A CN101169738A CNA2007100506314A CN200710050631A CN101169738A CN 101169738 A CN101169738 A CN 101169738A CN A2007100506314 A CNA2007100506314 A CN A2007100506314A CN 200710050631 A CN200710050631 A CN 200710050631A CN 101169738 A CN101169738 A CN 101169738A
- Authority
- CN
- China
- Prior art keywords
- task
- node
- formation
- information
- processing module
- 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
Images
Landscapes
- Information Transfer Between Computers (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method for allocating a task during media content creation, belonging to the technical field of task allocation in a distributed type parallel system. The inventive method comprises rapidly determining a reasonable task allocation initial solution by a detection algorithm, selecting proper nodes capable of finishing the task, and reducing system variation to meet stability requirement; and flexibly adjusting a task allocation strategy by controlling a task queue allocation scheme to meet the quick-response requirement to the system under multivariate environment. The inventive task allocation method can make the system under dynamic environment to rapidly enter stable sate, so as to perform reasonable task allocation. The invention has good practicability and expansibility, not only meets the task allocation requirement in media content creation field, but also meets the task allocation requirement in other fields.
Description
Technical field
The present invention relates to a kind of technology of allocating task, particularly a kind of in the media content production process method of allocating task.
Background technology
Since computing machine and internet birth, it has developed by leaps and bounds to influencing one of the most extensive, that growth is the fastest, market potential is maximum industry in the world today, and being called is the industrial revolution for the third time---the outstanding feature of era of knowledge-driven economy.The miniaturization of computing machine and the raising of the network bandwidth make computer application that the development of making rapid progress take place, and the medium creation has also obtained performance more and more fully under this platform.Yet, in the practical application of media content creation, some problems have also appearred, and concrete manifestation is as follows: 1, Task Distribution is unbalanced.The medium production process is the process of a collaborative work, is cooperation and interactional process as the work between editor, art designing, director, cameraman, animation personnel, the equipment management personnel.All kinds of personnel also are the colonies of multi agent cooperation, and each one processing power has a lot of differences in the community of interest.If Task Distribution is unreasonable in this working chain, as the execution inefficiency of some links, and execute the task when very heavy, just had a strong impact on the carrying out of next step work.2, resource environment constantly changes: carry out the unit of medium creation, as animation, film company, its resource environment is in continuous variation, and the personnel that mainly show change, the buying and selling of hardware resource.Existing method for allocating tasks is if guaranteed the Task Distribution of system, optimization, the variation that can not conform well just; On the contrary, if the system that guaranteed is flexible and changeable, but can not allow system enter stable Task Distribution state.The distribute media content process particularly big for scale, that resource class is many does not also have ripe adaptive task distribution method at present.
Summary of the invention
The objective of the invention is to overcome the deficiencies in the prior art, provide a kind of distributed processing system that is in a plurality of nodes that possess processor respectively under the dynamic environment that makes to enter the technical method that a kind of steady state (SS) is carried out rational Task Distribution very soon.
A kind of adaptive media content creative objective assignment method in the distributed processing system that comprises a plurality of nodes that possess processor respectively, reasonably is assigned to each node with task and handles, and may further comprise the steps:
Mission bit stream processing module, node information processing module, duty mapping processing module, Task Distribution processing module are set in system;
Handle the mission bit stream that arrives by the mission bit stream processing module, set up the information of task queue, logger task formation, the detection times of logger task formation;
Handle the employee information that arrives by the node information processing module, set up node formation, record node queuing message;
Extract the information of task queue, node formation transmission by the duty mapping module, set up the duty mapping table;
By the Task Distribution processing module Task Distribution is handled to each node: the formation that selection has task to carry out, extract file leader's task, the information of index duty mapping table, obtain the node that to finish this task in the node formation, select the allocation scheme of task queue according to the detection times of task queue, Task Distribution is handled to the node that is fit to: if the detection times of task queue is less than the number of times of default, then select to survey allocation scheme, assigning the task to all nodes that check out handles, if the detection times of task queue is more than or equal to the number of times of default, then select the self-adaptation allocation scheme, result according to the query task mapping table, in conjunction with the information change situation and the free time thereof of node, assign the task to suitable node automatically and handle; After finishing the work, node gives the Task Distribution processing module with information feedback, the Task Distribution processing module is handled node and task queue according to feedack, the information of node and task queue is changed, and give node information processing module and mission bit stream processing module the information feedback that changes.
Described adaptive media content creative objective assignment method process flow diagram as shown in Figure 1.
The distributed processing system synoptic diagram of the described a plurality of nodes that possess processor respectively as shown in Figure 2, this system operates in the distributed network, network edge is the work node, the center is a server.Described work node is a processor, comprises the people and the processor of Operation Processor, can finish specific task.Described server comprises ftp server, the WEB server, Task Distribution server etc., described mission bit stream processing module, node information processing module, duty mapping processing module, Task Distribution processing module all are operated on the Task Distribution server, realize the function of management work node, allocating task.
Described mission bit stream processing module is carried out following the processing: set up information, the logger task formation detection times of task queue, logger task formation, may further comprise the steps:
Step S301: new task arrives, for new task is distributed a mission bit stream storage element, typing mission bit stream;
Step S302: task is classified, describe according to mission bit stream new task is included into satisfactory task queue:
Step S321: initial selected file leader's formation, if the information of this formation and new task relatively identical then the requirement of joining the team is satisfied in explanation, proceeds to step S304, if inequality then enter into step S322,
Step S322: select next formation, if the information of this formation and new task relatively identical then proceed to step S304, if difference and current queue just proceed to step S303 for last formation, otherwise is proceeded step S322;
Step S303: newly-built task queue, the initialization task formation, the detection times (Count) that task queue is set is 0 (detection times refers to the number of times of finishing with generic task here), and queuing message is sent to the duty mapping processing module;
Step S304: new task is added this formation, extract the information that the node information processing module transmits, if the node formation has new node to add, the detection times that task queue then is set is 0; If no new node adds, then according to Task Distribution processing module feedack, the detection times of record queue finishes.The mission bit stream treatment scheme as shown in Figure 3.
Mission bit stream described here can comprise urgent degree of task code name, task description, task resource demand, task time, ability to work requirement; Described task queue essence is the single linked list of mission bit stream storage unit, and data structure definition comprises: queue number, and team's head pointer, formation is long, allocation scheme (Allocation), task description, counting (Count) is surveyed in the ability to work requirement.The synoptic diagram of task queue as shown in Figure 4.
Described node information processing module is carried out following the processing: set up the information of node formation, the formation of record node, may further comprise the steps:
Step S501: new node adds, for new node distributes a node information storage element, typing node information;
Step S502: carry out the node classification, describe according to node information new node is included into satisfactory node formation:
Step S521: initial selected file leader's formation, if the information of this formation and new node relatively identical then the requirement of joining the team is satisfied in explanation, proceeds to step S504, if inequality then enter into step 522,
Step S522: select next formation, if the information of this formation and new node relatively identical then proceed to step S504, if difference and current queue just proceed to step S503 for last formation, otherwise is proceeded step S522;
Step S503: newly-built node formation is sent to the duty mapping processing module with information;
Step S504: new node is added this formation, information is sent to the mission bit stream processing module, finish.The node information treatment scheme as shown in Figure 5.
Described node information can comprise node code name, ability to work, idle degree, ability to work requirement; Node information memory cell data organization definition is as follows, node number, and the ability to work requirement, ability to work (Ability), idle degree (Wait), the node queue data structure is defined as follows, queue number, team's head pointer, formation is long, the ability to work requirement.The synoptic diagram of node formation as shown in Figure 6.
Described duty mapping processing module is carried out following the processing: the information of extracting mission bit stream processing module, the transmission of node information processing module, according to identical information description, for example ability to work requires identical, set up the corresponding duty mapping table of task queue and node formation, task queue and node formation are mapped, for Task Distribution provides foundation, may further comprise the steps:
Step S701: set up mapping condition and note down according to task queue or node queuing message, set up the duty mapping table;
Step S702: the information of extracting node formation or task queue;
Step S703:, in the task mapping table, seek the record that conforms to mapping condition according to the described content of queuing message:
Step S731: choose the first trip record of duty mapping table, relatively if the information of formation and record identical, is carried out step S704, otherwise is carried out step S732,
Step S732: choose next record, relatively if the information of formation and record identical, is carried out step S704, if difference and current record are last, just set up the new record item and proceeds to step S704, otherwise proceed step S732;
Step S704: in record, deposit queuing message in according to the queue type among the step S702.The treatment scheme of duty mapping as shown in Figure 7.
Described Task Distribution processing module is carried out following the processing:
Step S801: file leader's task queue is selected in initialization, judges whether current queue has task to need execution, if having, then carries out step S803, if not then carry out step S802;
Step S802: select the next task formation, judge whether current queue has task to need execution, if having, then carries out step S803, if not then proceed step S802;
Step S803: extract file leader's task;
Step S804: index duty mapping table, promptly describe and find corresponding node formation according to mission bit stream, obtain that all can finish the node of this task in the node formation;
Step S805: whether Task Distribution, the detection times of judging task queue, are handled Task Distribution according to the allocation scheme of judged result selection task queue less than default value to the node that is fit to:
If the detection times of task queue less than the number of times of default, is then selected to survey allocation scheme, assign the task to all nodes that check out and handle,
If the detection times of task queue is more than or equal to the number of times of default, then select the self-adaptation allocation scheme, according to the result of query task mapping table,, assign the task to suitable node automatically and handle in conjunction with the information change situation and the free time thereof of node;
After finishing the work, node gives the Task Distribution processing module with information feedback, the Task Distribution processing module is handled node and task queue according to feedack, the information of node and task queue is changed, and give node information processing module and mission bit stream processing module the information feedback that changes.The treatment scheme of Task Distribution as shown in Figure 8.
Described node feedack comprises that task finishes the information of reflection such as used time, resource cost, works quality task performances, and described Task Distribution processing module is carried out following processing according to feedack to node and task queue:
Wait value (free time value) and Ability value (ability to work value) to node are handled: node be assigned to task then the Wait value increase, node finishes the work then that the Wait value reduces, node is unallocated to task then the Wait value do not change; Node finish the work the used time few, resource cost is little, the works quality is good, then the Ability value of node obtains increasing, on the contrary then the Ability value of node is reduced.Can carry out rewards and punishments according to predefined rewards and punishments function, such as the Ability value of determining how to change node by simple rewards and punishments computing function: Beta=task completeness * a1+ resource cost * a2+ works are estimated * a3, a1+a2+a3=1 wherein, the high more expression task of Beta is finished well more, the Beta of all node correspondences relatively, if the highest awarding, following change: Ability=Ability+1 takes place in node Ability value; IfAbility>=MAX then Ability=MAX; Otherwise punish being changed to of node Ability value: Ability=Ability-1; If Ability<=1 then Ability=1; Ability represents that node carries out the ability to work of this task, and high more expression executive capability is strong more.
Detection times to task queue is handled: the detection times (Count) of the back task queue of finishing the work adds one.According to the predefined detection times of system, such as T time, the detection times of task queue is compared with it, if Count>=T, represent that this task queue distributes fully through T time after, understood the ability to work of each node, can enter the self-adaptation distribution.The T value can be set according to the needs of system, and T is big more, and it is slow more that system enters the self-adaptation state, enters the more stable optimization of back system.Otherwise it is fast more that system enters the self-adaptation state, enters the back system and optimize be not as stable as the former.
The situation of change of task queue is fed back to the mission bit stream processing module, the situation of change of node is fed back to the node information processing module.
Described self-adaptation allocation scheme is that system extracts the Ability value and the Wait value of each node, computing function Alpha value, Alpha represents the value of the adaptive task of node, assign the task to the maximum node of Alpha value, such as Alpha=Ability*a1+Wait*a2, a1+a2=1 wherein.Alpha is directly proportional with ability to work, is inversely proportional to suitable more this task of the big more expression of Alpha with idle degree.Just can realize the automatic search of system by this principle distribution, and, carry out Task Distribution adaptively, can make system reach optimum or approaching optimum state soon with reference to the node idle condition to the node ability to work.
Compared with prior art, the invention has the beneficial effects as follows: the present invention determines rational Task Distribution initial solution apace by probe algorithm, and the suitable node that selection can be finished the work allows system reduce shake, satisfies the requirement of stability; And, adjust the Task Distribution strategy neatly by control to the task queue allocation scheme, satisfy system's rapid-action requirement under the changeable situation of environment.Method for allocating tasks provided by the present invention can make the system that is under the dynamic environment enter a kind of steady state (SS) very soon, carries out rational Task Distribution.The present invention has good practicality and extendability, is not only applicable to the needs of media content creation field Task Distribution, and can satisfy the needs of other field Task Distribution.
Description of drawings
Fig. 1 is adaptive media content creative objective assignment method process flow diagram.
Fig. 2 is adaptive media content creation network structure.
Fig. 3 is the mission bit stream processing flow chart.
Fig. 4 is a multitask formation synoptic diagram.
Fig. 5 is the node information processing flow chart.
Fig. 6 is many nodes formation synoptic diagram.
Fig. 7 is the duty mapping processing flow chart.
Fig. 8 is the Task Distribution process flow diagram.
Fig. 9 is the three-dimensional animation production process flow diagram.
Embodiment
The present invention is described in further detail below in conjunction with accompanying drawing and preferred embodiment, but this and do not mean that limiting the scope of the invention.
The preferred embodiments of the present invention relate to the Task Distribution in the media content production process.The practical application of media content creation mainly comprises: cartoon making, production of film and TV, development of games.Further classification comprises material creation, role's modeling, musical composition, storyboard creation, scene rendering or the like.The characteristics of media content creation have: task is various, and the degree of association is big, personnel's complexity.Therefore the creation of the media content of a success must be coordinated the relation between the staff, carries out rational Task Distribution.Below with media content creation application---the three-dimensional character cartoon making is an example, and Task Distribution is analyzed.The three-dimensional animation production flow process as shown in Figure 9.
Method for allocating tasks in the three-dimensional character animation process comprises: mission bit stream processing module, node information processing module, duty mapping processing module and Task Distribution processing module are set in system;
As shown in Figure 3, described mission bit stream processing module is carried out following the processing: set up information, the logger task formation detection times of task queue, logger task formation, may further comprise the steps:
Step S301: new task arrives, for new task is distributed a mission bit stream storage element, and the typing mission bit stream, the mission bit stream of three-dimensional character animation process is described and is seen Table 1;
Step S302: task is classified, and the main task classification of three-dimensional character animation process sees Table 1, describes according to mission bit stream new task is included into satisfactory task queue:
Step S321: initial selected file leader's formation, promptly at first select material creation formation, the information of this formation and new task relatively, if identical then the satisfied requirement of joining the team of explanation proceeds to step S304, if inequality then enter into step S322,
Step S322: select next formation, if the information of this formation and new task relatively identical then proceed to step S304, if difference and current queue just proceed to step S303 for last formation, otherwise is proceeded step S322;
Step S303: newly-built task queue, the initialization task formation, the detection times that task queue is set is 0, and queuing message is sent to the duty mapping resume module;
Step S304: new task is added this formation, extract the information that the node information processing module transmits, if the node formation has new node to add, the detection times that task queue then is set is 0; If no new node adds, then according to Task Distribution processing module feedack, the detection times of record queue finishes.
Mission bit stream described here comprises urgent degree of task code name, task description, task resource demand, task time, ability to work requirement; Described task queue essence is the single linked list of mission bit stream storage unit, and data structure definition comprises: queue number, and team's head pointer, formation is long, allocation scheme (Allocation), task description, counting (Count) is surveyed in the ability to work requirement.The task queue synoptic diagram as shown in Figure 4.
Table 1 three-dimensional character cartoon making mission statement table
Task queue | Task description | The ability to work demand | The node formation |
The material creation | Give lively character of surface machine-made model | Be familiar with the color and the texture of various objects | Material creator |
Modeling | The object that occurs in the role that appears on the scene and the scene is carried out modeling, move modeling as 3D role's modeling, 3D. | Be familiar with the relative merits of various modeling pattern, can select different modeling methods to carry out modeling according to different needs. | The modeling personnel |
The audio creation | Light is complementary with the atmosphere of plot, strengthens the effect and the drama of film | Professional music knowledge | Art director's sound effects man |
Light is provided with | Light is complementary with the atmosphere of plot, strengthens the effect and the drama of film | Be familiar with various light effects | Design personnel |
Text editing | Invention is strong, and the literal grounding in basic skills is deep | The text editing personnel | |
... | ... | ... | ... |
Play up | Allow the picture of playing up out that the sense of reality is arranged | Be familiar with various rendering effect, as oil paint effect, flat effect, the color plumbous effect of being coated with of two dimension.Can select diverse ways to play up according to different needs. | Play up personnel |
As shown in Figure 5, described node information processing module is carried out following the processing: set up the information of node formation, the formation of record node, may further comprise the steps:
Step S501: new node adds, for new node distributes a node information storage element, and the typing node information, the required node technical ability of three parallactic angle look cartoon making requires to see Table 1;
Step S502: carry out the node classification, describe according to node information new node is included into satisfactory node formation, the node formation is described and is seen Table 1:
Step S521: initial selected file leader's formation, i.e. material creation formation, if the information of this formation and new node relatively identical then the requirement of joining the team is satisfied in explanation, proceeds to step S504, if inequality then enter into step 522,
Step S522: select next formation, if the information of this formation and new node relatively identical then proceed to step S504, if difference and current queue just proceed to step S503 for last formation, otherwise is proceeded step S522;
Step S503: newly-built node formation, initialization node queuing data is provided with queue number, information is sent to the duty mapping processing module handles;
Step S504: new node is added this formation, send information to the mission bit stream processing module, finish.
Described node information comprises node code name, ability to work, idle degree, ability to work requirement; Node information memory cell data organization definition is as follows, node number, and the ability to work requirement, ability to work (Ability), idle degree (Wait), the node queue data structure is defined as follows, queue number, team's head pointer, formation is long, the ability to work requirement.Node formation synoptic diagram as shown in Figure 6.
As shown in Figure 7, described duty mapping processing module is carried out following the processing: the information that receives mission bit stream processing module, the transmission of node information processing module, according to identical information description, for example ability to work requires identical, set up the corresponding duty mapping table of task queue and node formation, task queue and node formation are mapped,, may further comprise the steps for Task Distribution provides foundation:
Step S701: set up mapping condition and note down according to task queue or node queuing message, set up the duty mapping table, see Table 2;
Step S702: the information of extracting node formation or task queue;
Step S703:, in the task mapping table, seek the record that conforms to mapping condition according to the described content of queuing message;
Step S731: choose the first trip record of duty mapping table, relatively if the information of formation and record identical, is carried out step S704, otherwise is carried out step S732;
Step S732: choose next record, relatively if the information of formation and record identical, is carried out step S704; If difference and current record are last, just set up the new record item and proceed to step S704, otherwise proceed step S732;
Step S704: in record, deposit queuing message in according to the queue type among the step S702.
Table 2 duty mapping table
Recording mechanism | Task queue | The node formation | Mapping condition | ... |
Record 1 | Task queue a | Node formation b | Ability to work requires to be equal to | ... |
Record 2 | Task queue b | Node formation r node formation t | Ability to work requires to be equal to | ... |
... | ... | ... | ... | ... |
As shown in Figure 8, described Task Distribution processing module is carried out following the processing:
Step S801: file leader's task queue is selected in initialization, promptly at first selects material creation formation, judges whether current queue has task to need execution, if having, then carries out step S803, if not then carry out step S802;
Step S802: select the next task formation, judge whether current queue has task to need execution, if having, then carries out step S803, if not then proceed step S802;
Step S803: extract file leader's task;
Step S804: index duty mapping table, promptly describe and find corresponding node formation according to mission bit stream, obtain that all can finish the node of this task in the node formation;
Step S805: whether Task Distribution, the detection times of judging task queue, are handled Task Distribution according to the allocation scheme of judged result selection task queue less than default value to the node that is fit to:
If the detection times of task queue less than the number of times of default, is then selected to survey allocation scheme, assign the task to all nodes that check out and handle;
If the detection times of task queue is more than or equal to the number of times of default, then select the self-adaptation allocation scheme, according to the result of query task mapping table,, assign the task to suitable node automatically and handle in conjunction with the information change situation and the free time thereof of node;
After finishing the work, node gives the Task Distribution processing module with information feedback, the Task Distribution processing module is handled node and task queue according to feedack, the information of node and task queue is changed, and give node information processing module and mission bit stream processing module the information feedback that changes.
Described node feedack comprises that task finishes the information of reflection such as used time, resource cost, works quality task performances, and described Task Distribution processing module is carried out following processing according to feedack to node and task queue:
Wait value (free time value) and Ability value (ability to work value) to node are handled: node be assigned to task then the Wait value increase, node finishes the work then that the Wait value reduces, unallocated to task then the Wait value do not change; Node finish the work the used time few, resource cost is little, the works quality is good, then the Ability value of node obtains increasing, on the contrary then the Ability value of node is reduced.Can carry out rewards and punishments according to predefined rewards and punishments function, such as the Ability value of determining how to change node by simple rewards and punishments computing function: Beta=task completeness * a1+ resource cost * a2+ works are estimated * a3, a1+a2+a3=1 wherein, the high more expression task of Beta is finished well more, the Beta of all node correspondences relatively, if the highest awarding, following change: Ability=Ability+1 takes place in node Ability value; IfAbility>=MAX then Ability=MAX; Otherwise punish being changed to of node Ability value: Ability=Ability-1; If Ability<=1 then Ability=1; Ability represents that node carries out the ability to work of this task, and high more expression executive capability is strong more.
Detection times to task queue is handled: the detection times (Count) of the back task queue of finishing the work adds one.According to the predefined detection times of system, such as T time, the detection times of task queue is compared with it, if Count>=T, represent that this task queue distributes fully through T time after, understood the ability to work of each node, can enter the self-adaptation distribution.T is big more, and it is slow more that system enters the self-adaptation state, enters the more stable optimization of back system.Otherwise it is fast more that system enters the self-adaptation state, enters the back system and optimize be not as stable as the former.
The situation of change of task queue is fed back to the mission bit stream processing module, the situation of change of node is fed back to the node information processing module.
Described self-adaptation allocation scheme is that system extracts the Ability value and the Wait value of each node, computing function Alpha value, Alpha represents the value of the adaptive task of node, assign the task to the maximum node of Alpha value, such as Alpha=Ability*a1+Wait*a2, a1+a2=1 wherein.Alpha is directly proportional with ability to work, is inversely proportional to suitable more this task of the big more expression of Alpha with idle degree.Just can realize the automatic search of system by this principle distribution, and, carry out Task Distribution adaptively, can make system reach optimum or approaching optimum state soon with reference to the node idle condition to the node ability to work.
Task two states may occur after treatment: 1. optimum state, and the people that ability to work is the strongest, free time is more is distributed in all subtasks, and overall task is finished.2. suboptimum state, parton Task Distribution give that ability to work is strong, more people of free time, and overall task is finished.Two states all is to occur under the situation that the assurance overall work can be finished, and it is minimum that the former can allow system resource expend, system optimal; It is less that the latter can allow system resource expend, and system is more excellent.Both sides relation: Task Distribution can not reach optimum state, then transfers suboptimum to.
Claims (8)
1. an adaptive media content creative objective assignment method in the distributed processing system that comprises a plurality of nodes that possess processor respectively, reasonably is assigned to each node with task and handles, and it is characterized in that may further comprise the steps:
Mission bit stream processing module, node information processing module, duty mapping processing module, Task Distribution processing module are set in system;
Handle the mission bit stream that arrives by the mission bit stream processing module, set up the information of task queue, logger task formation, the detection times of logger task formation;
Handle the employee information that arrives by the node information processing module, set up node formation, record node queuing message;
Extract the information of task queue, node formation transmission by the duty mapping module, set up the duty mapping table;
By the Task Distribution processing module Task Distribution is handled to each node: the formation that selection has task to carry out, extract file leader's task, the information of index duty mapping table, obtain the node that to finish this task in the node formation, select the allocation scheme of task queue according to the detection times of task queue, Task Distribution is handled to the node that is fit to: if the detection times of task queue is less than the number of times of default, then select to survey allocation scheme, assigning the task to all nodes that check out handles, if the detection times of task queue is more than or equal to the number of times of default, then select the self-adaptation allocation scheme, result according to the query task mapping table, in conjunction with the information change situation and the free time thereof of node, assign the task to suitable node automatically and handle; After finishing the work, node gives the Task Distribution processing module with information feedback, the Task Distribution processing module is handled node and task queue according to feedack, the information of node and task queue is changed, and give node information processing module and mission bit stream processing module the information feedback that changes.
2. adaptive media content creative objective assignment method according to claim 1, it is characterized in that following processing of described mission bit stream processing module execution: set up information, the logger task formation detection times of task queue, logger task formation, may further comprise the steps:
Step S301: new task arrives, for new task is distributed a mission bit stream storage element, typing mission bit stream;
Step S302: task is classified, describe according to mission bit stream new task is included into satisfactory task queue:
Step S321: initial selected file leader's formation, if the information of this formation and new task relatively identical then the requirement of joining the team is satisfied in explanation, proceeds to step S304, if inequality then enter into step S322,
Step S322: select next formation, if the information of this formation and new task relatively identical then proceed to step S304, if difference and current queue just proceed to step S303 for last formation, otherwise is proceeded step S322;
Step S303: newly-built task queue, the initialization task formation, the detection times that task queue is set is 0, and queuing message is sent to the duty mapping processing module;
Step S304: new task is added this formation, extract the information that the node information processing module transmits, if the node formation has new node to add, the detection times that task queue then is set is 0, if no new node adds, then according to the information of Task Distribution processing module transmission, the detection times of record queue finishes.
3. adaptive media content creative objective assignment method according to claim 1 is characterized in that following processing of described node information processing module execution: set up the information of node formation, the formation of record node, may further comprise the steps:
Step S501: new node adds, for new node distributes a node information storage element, typing node information;
Step S502: node is classified, describe according to node information new node is included into satisfactory node formation:
Step S521: initial selected file leader's formation, if the information of this formation and new node relatively identical then the requirement of joining the team is satisfied in explanation, proceeds to step S504, if inequality then enter into step 522,
Step S522: select next formation, if the information of this formation and new node relatively identical then proceed to step S504, if difference and current queue just proceed to step S503 for last formation, otherwise is proceeded step S522;
Step S503: newly-built node formation is sent to the duty mapping processing module with information;
Step S504: new node is added this formation, information is sent to the mission bit stream processing module, finish.
4. adaptive media content creative objective assignment method according to claim 1, it is characterized in that following processing of described duty mapping processing module execution: the information of extracting mission bit stream processing module, the transmission of node information processing module, according to identical information description, set up the corresponding duty mapping table of task queue and node formation, task queue and node formation are mapped, for Task Distribution provides foundation, may further comprise the steps:
Step S701: set up mapping condition and note down according to task queue or node queuing message, set up the duty mapping table;
Step S702: the information of extracting node formation or task queue;
Step S703:, in the task mapping table, seek the record that conforms to mapping condition according to the described content of queuing message:
Step S731: choose the first trip record of duty mapping table, relatively if the information of formation and record identical, is carried out step S704, otherwise is carried out step S732,
Step S732: choose next record, relatively if the information of formation and record identical, is carried out step S704, if difference and current record are last, just set up the new record item and proceeds to step S704, otherwise proceed step S732;
Step S704: in this record, deposit queuing message in according to the queue type among the step S702.
5. adaptive media content creative objective assignment method according to claim 1 is characterized in that following processing of described Task Distribution processing module execution:
Step S801: file leader's task queue is selected in initialization, judges whether current queue has task to need execution, if having, then carries out step S803, if not then carry out step S802;
Step S802: select the next task formation, judge whether current queue has task to need execution, if having, then carries out step S803, if not then proceed step S802;
Step S803: extract file leader's task;
Step S804: index duty mapping table, promptly describe and find corresponding node formation according to mission bit stream, obtain that all can finish the node of this task in the node formation;
Step S805: whether Task Distribution, the detection times of judging task queue, are handled Task Distribution according to the allocation scheme of judged result selection task queue less than default value to the node that is fit to:
If the detection times of task queue less than the number of times of default, is then selected to survey allocation scheme, assign the task to all nodes that check out and handle,
If the detection times of task queue is more than or equal to the number of times of default, then select the self-adaptation allocation scheme, according to the result of query task mapping table,, assign the task to suitable node automatically and handle in conjunction with the information change situation and the free time thereof of node;
After finishing the work, node gives the Task Distribution processing module with information feedback, the Task Distribution processing module is handled node and task queue according to feedack, the information of node and task queue is changed, and give node information processing module and mission bit stream processing module the information feedback that changes.
6. adaptive media content creative objective assignment method according to claim 5 is characterized in that described Task Distribution processing module carries out following processing according to feedack to node and task queue:
Wait value and Ability value to node are handled: node be assigned to task then the Wait value increase, node finishes the work then that the Wait value reduces, node is unallocated to task then the Wait value do not change; Node finish the work the used time few, resource cost is little, the works quality is good, then the Ability value of node obtains increasing, on the contrary then the Ability value of node is reduced;
Detection times to task queue is handled: the detection times of the back task queue of finishing the work adds one;
The situation of change of node is fed back to the node information processing module, the situation of change of task queue is fed back to the mission bit stream processing module.
7. adaptive media content creative objective assignment method according to claim 6, it is characterized in that described Task Distribution processing module carries out rewards and punishments according to predefined rewards and punishments function the Ability value of node is handled, described rewards and punishments function is: Beta=task completeness * a1+ resource cost * a2+ works are estimated * a3, a1+a2+a3=1 wherein, Beta represents the task performance, the Beta of all node correspondences relatively, if the highest then award, following change: Ability=Ability+1 takes place in node Ability value; IfAbility>=MAX then Ability=MAX; Otherwise punish being changed to of node Ability value: Ability=Ability-1; If Ability<=1 then Ability=1.
8. adaptive media content creative objective assignment method according to claim 5, it is characterized in that described self-adaptation allocation scheme is, the Task Distribution processing module is extracted the Ability value and the Wait value of each node, computing function Alpha value, assign the task to the maximum node of Alpha value, described Alpha=Ability*a1+Wait*a2, wherein a1+a2=1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2007100506314A CN100483353C (en) | 2007-11-28 | 2007-11-28 | Self-adaptive media content creative objective assignment method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2007100506314A CN100483353C (en) | 2007-11-28 | 2007-11-28 | Self-adaptive media content creative objective assignment method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101169738A true CN101169738A (en) | 2008-04-30 |
CN100483353C CN100483353C (en) | 2009-04-29 |
Family
ID=39390374
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB2007100506314A Expired - Fee Related CN100483353C (en) | 2007-11-28 | 2007-11-28 | Self-adaptive media content creative objective assignment method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100483353C (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102208222A (en) * | 2011-03-29 | 2011-10-05 | 中广核工程有限公司 | Control method and system thereof of nuclear power plant operation |
CN102938127A (en) * | 2012-11-29 | 2013-02-20 | 北京奇虎科技有限公司 | Method and device for providing information |
CN103019649A (en) * | 2012-11-29 | 2013-04-03 | 北京奇虎科技有限公司 | Information providing method and device |
CN103377076A (en) * | 2012-04-28 | 2013-10-30 | 国际商业机器公司 | Method and system for adjusting task execution plans during operation |
WO2015014150A1 (en) * | 2013-08-01 | 2015-02-05 | Tencent Technology (Shenzhen) Company Limited | Method,apparatus,and application platform for updating application object attribute |
WO2017128933A1 (en) * | 2016-01-29 | 2017-08-03 | 深圳市艾特大师网络科技有限公司 | Processing method, device, and system for work |
WO2017128920A1 (en) * | 2016-01-29 | 2017-08-03 | 深圳市艾特大师网络科技有限公司 | Method, device, and terminal for message processing |
CN108234565A (en) * | 2016-12-21 | 2018-06-29 | 天脉聚源(北京)科技有限公司 | A kind of method and system of server cluster processing task |
CN108921426A (en) * | 2018-06-28 | 2018-11-30 | 景致三维(江苏)股份有限公司 | Three-dimensional modeling method and system |
CN111275276A (en) * | 2018-12-05 | 2020-06-12 | 阿里巴巴集团控股有限公司 | Video creation task management method and device and electronic equipment |
-
2007
- 2007-11-28 CN CNB2007100506314A patent/CN100483353C/en not_active Expired - Fee Related
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102208222A (en) * | 2011-03-29 | 2011-10-05 | 中广核工程有限公司 | Control method and system thereof of nuclear power plant operation |
CN102208222B (en) * | 2011-03-29 | 2013-09-11 | 中广核工程有限公司 | Control method and system thereof of nuclear power plant operation |
CN103377076A (en) * | 2012-04-28 | 2013-10-30 | 国际商业机器公司 | Method and system for adjusting task execution plans during operation |
US9690617B2 (en) | 2012-04-28 | 2017-06-27 | International Business Machines Corporation | Adjustment of a task execution plan at runtime |
CN102938127B (en) * | 2012-11-29 | 2016-07-20 | 北京奇虎科技有限公司 | A kind of method and apparatus for providing information |
CN103019649A (en) * | 2012-11-29 | 2013-04-03 | 北京奇虎科技有限公司 | Information providing method and device |
CN103019649B (en) * | 2012-11-29 | 2016-08-31 | 北京奇虎科技有限公司 | Information providing method and equipment |
CN102938127A (en) * | 2012-11-29 | 2013-02-20 | 北京奇虎科技有限公司 | Method and device for providing information |
WO2015014150A1 (en) * | 2013-08-01 | 2015-02-05 | Tencent Technology (Shenzhen) Company Limited | Method,apparatus,and application platform for updating application object attribute |
US10204125B2 (en) | 2013-08-01 | 2019-02-12 | Tencent Technology (Shenzhen) Company Limited | Method, apparatus, and application platform for updating application object attribute |
WO2017128933A1 (en) * | 2016-01-29 | 2017-08-03 | 深圳市艾特大师网络科技有限公司 | Processing method, device, and system for work |
WO2017128920A1 (en) * | 2016-01-29 | 2017-08-03 | 深圳市艾特大师网络科技有限公司 | Method, device, and terminal for message processing |
CN108234565A (en) * | 2016-12-21 | 2018-06-29 | 天脉聚源(北京)科技有限公司 | A kind of method and system of server cluster processing task |
CN108921426A (en) * | 2018-06-28 | 2018-11-30 | 景致三维(江苏)股份有限公司 | Three-dimensional modeling method and system |
CN111275276A (en) * | 2018-12-05 | 2020-06-12 | 阿里巴巴集团控股有限公司 | Video creation task management method and device and electronic equipment |
Also Published As
Publication number | Publication date |
---|---|
CN100483353C (en) | 2009-04-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN100483353C (en) | Self-adaptive media content creative objective assignment method | |
CN102881046B (en) | Method for generating three-dimensional electronic map | |
CN106875320B (en) | Efficient visual analysis method for ship navigation data in cloud environment | |
Li et al. | Application of virtual reality technology and digital twin in digital media communication | |
Shukla et al. | FAT-ETO: Fuzzy-AHP-TOPSIS-Based efficient task offloading algorithm for scientific workflows in heterogeneous fog–cloud environment | |
CN115293623A (en) | Training method and device for production scheduling model, electronic equipment and medium | |
CN108595882B (en) | Yinyi planning pipeline BIM design pipeline cost generation system | |
Besuievsky et al. | Customizable lod for procedural architecture | |
Hu et al. | Research on 3D animation character design based on multimedia interaction | |
Walton et al. | Ecosystems thinking and modern platform-based ecosystem theory | |
Moura | 3D density histograms for criteria-driven edge bundling | |
Le-Khac et al. | A new approach for distributed density based clustering on grid platform | |
Annette et al. | Ranking cloud render farm services for a multi criteria recommender system | |
Kumar | The The Red Queen Effect and How to Evade the Red Queen Effect by using Generative AI: Preparing Companies for Industry 5.0 | |
Pereira et al. | Using neuroevolution for predicting mobile marketing conversion | |
Wu | Interior Design Recommendation Technology Based on Interactive Genetic Algorithm | |
Ruiz et al. | Including dynamic adaptative topology to particle swarm optimization algorithms | |
JP6196265B2 (en) | Program, information processing apparatus and control method | |
Lang et al. | Design of Interactive Virtual System of Architectural Space Based on Multi-Objective Optimization Algorithm | |
Zhuge et al. | Augmented Reality Based on Network Physics and 6G for Immersive Experience of Digital Media Art | |
Alamsyah et al. | Digital Society Ecosystem Impact on Creative Industry | |
Đuričin et al. | Rejuvenation of business management tools in Industry 4.0 | |
Zelinka et al. | On Relation Between Swarm and Evolutionary Dynamics and Complex Networks | |
Zhang et al. | Task-load aware and predictive-based workflow scheduling in cloud-edge collaborative environment | |
Gomes et al. | Digital Twin of a Cloud Data Centre: An OpenStack Cluster Visualisation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090429 Termination date: 20191128 |