CN108446455A - A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm - Google Patents
A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm Download PDFInfo
- Publication number
- CN108446455A CN108446455A CN201810164337.4A CN201810164337A CN108446455A CN 108446455 A CN108446455 A CN 108446455A CN 201810164337 A CN201810164337 A CN 201810164337A CN 108446455 A CN108446455 A CN 108446455A
- Authority
- CN
- China
- Prior art keywords
- solution
- fireworks
- spark
- domination
- node
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2117/00—Details relating to the type or aim of the circuit design
- G06F2117/08—HW-SW co-design, e.g. HW-SW partitioning
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm:Random initializtion meets N number of solution of hardware area constraints, and N number of solution corresponds to N number of fireworks in fireworks algorithm;N number of solution is put into a set P, according to Pareto optimum theories, finds out the non-domination solution in set P successively, and it is 1,2 ... that the non-dominant grade for the non-domination solution found out successively, which is set gradually,;According to crowding computational methods, the dispersibility of the non-domination solution in same non-dominant grade is calculated;All non-domination solutions for obtaining non-dominant grade and dispersibility are ranked up;The number and fireworks explosion amplitude that fireworks explosion generates spark are calculated according to sequencing information;Generate all explosive sparks;Generate all Gauss sparks;All fireworks, explosive spark and Gauss spark are ranked up, the top n fireworks or explosive spark or Gauss spark obtained are as follow-on fireworks;Reach the iterations of setting.The present invention improves the speed of service of system and reduces the power consumption of system.
Description
Technical field
The present invention relates to a kind of Method for HW/SW partitioning.It is applied in complex embedded system design more particularly to one kind
The multiple target Method for HW/SW partitioning based on fireworks algorithm.
Background technology
1, fireworks algorithm
Fireworks algorithm is a kind of novel Swarm Intelligent Algorithm that invention is inspired by fireworks explosion phenomenon, main to use
In solving the problems, such as the mathematically higher NP-hard of complexity.The main thought of fireworks algorithm is:It randomly places on the ground several
Fireworks simultaneously assess their quality, and high-quality fireworks explosive emission goes out more spark, and is all centered around its week
It encloses, shows grand scene;The spark number that ropy fireworks explosive emission goes out is less and more dispersed.Correspond to number
Knowledge is inscribed:First several schemes solved the problems, such as are generated at random and these schemes are assessed.It is preferable for fitness
Solution generates more new solution in its vicinity, distributes more resource and carries out neighborhood search to it, mainly
For being exploited to solution space;It is fewer being generated apart from its remote position for the solution that fitness is poor
New solution is mainly used for detecting solution space.Exploitation and detectivity of the fireworks algorithm in balanced algorithm, and
Preferable performance is shown in terms of jumping out locally optimal solution.
2, Pareto optimum theories
In multi-objective optimization question, multiple optimization aims are often conflicting, preferably solve for one group in order to obtain
Certainly scheme, Italian economist's Pareto propose a kind of method of relatively multiple target solution.Pareto optimum theories it is basic
Concept is as follows:
Multi-objective optimization question:By taking minimization problem as an example, multi-objective optimization question can be described as following form:
min:
F (x)=(f1(x),f2(x),…,fn(x))T
subject to:
gi(x)≥0,i∈I
hj(x)=0, j ∈ E
Wherein f (x) is optimization aim, gi(x) and hj(x) it is respectively inequality constraints and equality constraint.
Pareto is dominated:One given object vector x=(x1,x2,…,xn) dominate another object vector y=(y1,
y2,…,yn) (be denoted as), and if only iff(xi)≤f(yi), andMake f
(xj)<f(yj)。
Pareto optimal solutions:If a solution x*It is referred to as Pareto optimal solutions, and if only if x*Not by others solution branch
Match.
Pareto optimal solution sets:Set XPareto=x | x is Pareto optimal solutions }.
The forward positions Pareto:Set FPareto=f (x) | x is Pareto optimal solutions }.
3, hardware-software partition
One complicated embedded system would generally be divided into multiple subtasks to realize, the realization method pair of subtask
The performance of entire embedded system has large effect.In hardware-software partition, it is assumed that by a complicated embedded system point
At N number of subtask, each subtask (can be indicated) with software realization (being indicated with 0) or with hardware realization with 1.Entirely
Splitting scheme can be indicated with one group of orderly binary sequence, such as 10110 indicate the 1st, 3,4 subtask software reality
It is existing, the 2nd, 5 subtask hardware realization.Subtask software or hardware realization at runtime, power consumption, in terms of area occupied
Show different performances.One group of short and low in energy consumption software and hardware of run time stroke on limited area occupied in order to obtain
Offshoot program, the present invention proposes a kind of multiple target Method for HW/SW partitioning based on fireworks algorithm, multiple to reach while optimize
Target obtains one group of solution of more excellent solution simultaneously with power consumption at runtime.The mathematics of multiple target hardware-software partition
Model can be expressed as:
min:
T=max TE (i) | 0<i<N}
subject to:
Wherein T, P and A indicate the total time of system operation, power consumption and area occupied.All subtasks for having dependence
A paths are formed, on one path, if having data dependence relation between two nodes and using software and hardware respectively
It realizes, then needing to calculate the call duration time between the node, TE (i) indicates the run time under the i-th paths.w(i)∈
{ 0,1 } indicate that the realization method of subtask, wherein w (i)=0 indicate that i-th of subtask software realization, w (i)=1 indicate i-th
A subtask hardware realization.ph(i) and ps(i) i-th of subtask hardware realization and the work(with software realization are indicated respectively
Consumption.ah(i) and as(i) i-th of subtask hardware realization and the area with software realization are indicated respectively.
Search capability is poor existing for existing multiple target Method for HW/SW partitioning, and efficiency is low, solves ropy problem.
Invention content
The technical problem to be solved by the invention is to provide one kind finding high-quality, number in acceptable time range
Amount is more, the multiple target Method for HW/SW partitioning based on fireworks algorithm of one group of big hardware-software partition scheme of difference.
The technical solution adopted in the present invention is:A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm, including
Following steps:
1) the N number of solution for meeting hardware area constraints of random initializtion, each solution by 0 and 1 form two into
Number processed, wherein each number represents a node, and all digits of binary number constitute total node, and N number of solution corresponds to fireworks
N number of fireworks in algorithm;
2) N number of solution is put into a set P, according to Pareto optimum theories, finds out the non-domination solution in set P,
The non-dominant grade for the non-domination solution found out is set as 1, and the non-domination solution is moved into from set P in another set Q;
Non-domination solution is found out, and non-dominant by what is found out for the second time according to Pareto optimum theories to all remaining solutions in set P
The non-dominant grade of solution is set as 2, equally moves into the non-domination solution found out for the second time in set Q from set P, and so on,
Until the element in set P is 0;
3) according to crowding computational methods, the dispersibility of the non-domination solution in same non-dominant grade is calculated;
4) all non-domination solutions for obtaining non-dominant grade and dispersibility are ranked up;
5) the number z that fireworks explosion generates spark is calculated according to sequencing informationiWith fireworks explosion amplitude Ai, calculation formula is such as
Under:
Wherein, zmax、zmin、AmaxAnd AminIt is pre-defined constant, zmaxAnd zminMaximum explosive spark number is indicated respectively
Mesh and minimum explosive spark number, AmaxAnd AminMaximum explosion amplitude and minimum explosion amplitude are indicated respectively;I indicates cigarette after sequence
Colored serial number;N indicates the number of fireworks;
6) explosive spark is generated, including:Z is generated to i-th of fireworksiA explosive spark obtains the position of explosive spark first
Set x=xi, xiIt indicates i-th of fireworks, a node is randomly choosed from total node number, if node is selected, when what is be selected
Node is 1 to be changed to 0, when selected node is 0 to be changed to 1, generates explosive spark, the calculating public affairs of wherein a node
Formula:A=AiRand (0,1), Ai are the explosion amplitudes of i-th of fireworks, and rand (0,1) is the random number between 0~1;
7) step 6) is repeated until generating all explosive sparks;
8) Gauss spark is generated, including:The position x of Gauss spark is obtained, x is randomly choosed from N number of fireworks, is saved to one
Point generates the random number of a Gaussian Profile, if the random number in setting range, when the node is 1 to be changed to 0, when
The node is 0 to be changed to 1, generates a Gauss spark;
9) step 8) is repeated until generating all Gauss sparks;
10) all fireworks, explosive spark and Gauss spark are ranked up according to step 2)-step 4), are obtained after sequence
The top n fireworks or explosive spark or Gauss spark obtained are as follow-on fireworks;
11) step 2)-step 8) is repeated until reaching the iterations of setting.
Crowding computational methods described in step 3) include:First to the non-domination solution in the same non-dominant grade according to
One optimization aim is ranked up, and is set any one non-dominant grade and is shared L non-domination solution, the 1st and l-th non-domination solution
Crowding be infinity, then k-th solution crowding using following formula calculating:
Crowd (k)=2 × ((t (k+1)-t (k-1))+(p (k-1)-p (k+1)))
The crowding of wherein crowd (k) k-th of solution of expression, 1<k<L, t and p are illustrated respectively in system in object space
It is total to execute time and total power consumption.
The standard of the step 4) sequence is:For any two solution m and n, when the non-dominant grade of m-th of solution is than n-th
The non-dominant grade of a solution it is low or m-th solution and n-th of solution in the same non-dominant grade and m-th solution crowding
The crowding solved than n-th is big, then better than n-th solution of m-th of solution.
Fireworks algorithm is applied to multiple target by a kind of multiple target Method for HW/SW partitioning based on fireworks algorithm of the present invention
The calculation formula of hardware-software partition problem, explosive spark number and explosion amplitude to fireworks algorithm is adjusted, can be effective
The search capability for solving to have algorithm is poor, and efficiency is low, solves ropy problem, is drawn to be efficiently completed multiple target software and hardware
Divide task.The method of the present invention can find high-quality in acceptable time range, and quantity is more, and difference big one group is soft or hard
Part splitting scheme realizes the division of its subtask on a processor to a complex embedded system, improves the operation speed of system
Degree and the power consumption for reducing system.
Specific implementation mode
A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm of the present invention is made in detail with reference to embodiment
It describes in detail bright.
A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm of the present invention, is to a complex embedded system
It realizes the division of its subtask on a processor, improve the speed of service of system and reduces the power consumption of system.
A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm of the present invention, includes the following steps:
1) the N number of solution for meeting hardware area constraints of random initializtion, each solution by 0 and 1 form two into
Number processed, wherein each number represents a node, and all digits of binary number constitute total node, and N number of solution corresponds to fireworks
N number of fireworks in algorithm;
2) N number of solution is put into a set P, according to Pareto optimum theories, finds out the non-domination solution in set P,
The non-dominant grade for the non-domination solution found out is set as 1, and the non-domination solution is moved into from set P in another set Q;
Non-domination solution is found out, and non-dominant by what is found out for the second time according to Pareto optimum theories to all remaining solutions in set P
The non-dominant grade of solution is set as 2, equally moves into the non-domination solution found out for the second time in set Q from set P, and so on,
Until the element in set P is 0;
3) according to crowding computational methods, the dispersibility of the non-domination solution in same non-dominant grade is calculated;
The crowding computational methods include:It is excellent according to one to the non-domination solution in the same non-dominant grade first
Change target (such as run time) to be ranked up, sets any one non-dominant grade and share L non-domination solution, the 1st and l-th
The crowding of non-domination solution is infinity, then the crowding of k-th of solution is calculated using following formula:Its mathematical meaning is with previous
A solution and value of the latter solution in object space are the perimeter of the rectangle of vertex composition.
Crowd (k)=2 × ((t (k+1)-t (k-1))+(p (k-1)-p (k+1)))
The crowding of wherein crowd (k) k-th of solution of expression, 1<k<L, t and p are illustrated respectively in system in object space
It is total to execute time and total power consumption.
4) all non-domination solutions for obtaining non-dominant grade and dispersibility are ranked up;
The standard of the sequence is:For any two solution m and n, solved than n-th when the non-dominant grade of m-th of solution
Non-dominant grade it is low or m-th solution and n-th of solution in the same non-dominant grade and m-th solution crowding ratio n-th
The crowding of a solution is big, then better than n-th solution of m-th of solution.
5) the number z that fireworks explosion generates spark is calculated according to sequencing informationiWith fireworks explosion amplitude Ai, calculation formula is such as
Under:
Wherein, zmax、zmin、AmaxAnd AminIt is pre-defined constant, zmaxAnd zminMaximum explosive spark number is indicated respectively
Mesh and minimum explosive spark number, AmaxAnd AminMaximum explosion amplitude and minimum explosion amplitude are indicated respectively;I indicates cigarette after sequence
Colored serial number;N indicates the number of fireworks;
6) explosive spark is generated, including:Z is generated to i-th of fireworksiA explosive spark obtains the position of explosive spark first
Set x=xi, xiIt indicates i-th of fireworks, a node is randomly choosed from total node number, if node is selected, when what is be selected
Node is 1 to be changed to 0, when selected node is 0 to be changed to 1, generates explosive spark, the calculating public affairs of wherein a node
Formula:A=AiRand (0,1), Ai are the explosion amplitudes of i-th of fireworks, and rand (0,1) is the random number between 0-1;
7) step 6) is repeated until generating all explosive sparks;
8) Gauss spark is generated, including:The position x of Gauss spark is obtained, x is randomly choosed from N number of fireworks, is saved to one
Point generates the random number of a Gaussian Profile, if the random number in setting range, when the node is 1 to be changed to 0, when
The node is 0 to be changed to 1, generates a Gauss spark;
9) step 8) is repeated until generating all Gauss sparks;
10) all fireworks, explosive spark and Gauss spark are ranked up according to step 2)-step 4), are obtained after sequence
The top n fireworks or explosive spark or Gauss spark obtained are as follow-on fireworks;
11) step 2)-step 8) is repeated until reaching the iterations of setting.
Preferred example is given below:
By taking 20 node hardware-software partition problems as an example, algorithm parameter setting:Maximum iteration MaxIter=1000,
Fireworks number FireNum=5, maximum explosive spark number zmax=10, minimum explosive spark number zmin=2, maximum explosion width
Spend Amax=10, minimum explosion amplitude Amin=2, Gauss spark number GaussNum=5.
1,5 solutions for meeting hardware area constraints, corresponding 5 fireworks are generated at random.Each fireworks by 20 two into
Array processed is at 0 represents software realization, and 1 represents hardware realization.Iterations iter=0.
2, fireworks are ranked up according to non-dominant grade and crowding.
3, the number z of fireworks explosive spark is calculated by following formula according to sequencing informationiWith explosion amplitude Ai。
Wherein zmax, zmin, AmaxAnd AminIt is pre-defined constant, indicates maximum explosive spark number respectively, it is minimum quick-fried
Fried spark number, maximum explosion amplitude and minimum explosion amplitude.I indicates that the serial number of fireworks after sequence, N indicate the number of fireworks.
4, explosive spark is generated, i-th of fireworks generates ziA spark.The position x=x of spark is initialized firsti, xiIt is i-th
The position of a fireworks.The interstitial content a=A of realization method will be changed by calculatingiRand (0,1), wherein rand (0,1) generate one
Random number between a 0 to 1.A node is randomly choosed from 20 nodes, if node is selected, which is become by 0
Become 0 for 1 or from 1.The 4th step is repeated until generating all sparks.
5,5 Gauss sparks are generated.One is randomly selected from 5 fireworks as initialization Gauss spark x, to the every of x
One generation, one random number randm (0,1), randm (0,1) are that mean value is 0, and variance is 1 to meet the random of Gaussian Profile
Number.If -0.5<randm(0,1)<0.5, then corresponding to position becomes 1 from 0 or becomes 0 from 1.It repeats the 5th step 5 times, generates 5
Gauss spark.
6, it sorts according to non-dominant grade and crowding to 5 fireworks, all explosive sparks and 5 Gauss sparks, and
Preceding 5 fireworks or explosive spark or Gauss spark are chosen as follow-on fireworks.Iterations iter adds 1.
7, step 2 to 6 is repeated until iterations reach maximum iteration iter=MaxIter.
8, export all non-domination solution, i.e., one group always to execute time and total power consumption as target, be about with area occupied
The multiple target hardware-software partition scheme of beam.
Claims (3)
1. a kind of multiple target Method for HW/SW partitioning based on fireworks algorithm, which is characterized in that include the following steps:
1) the N number of solution for meeting hardware area constraints of random initializtion, the binary number that each solution is made of 0 and 1,
Wherein, each number represents a node, and all digits of binary number constitute total node, and N number of solution corresponds to fireworks algorithm
In N number of fireworks;
2) N number of solution is put into a set P, according to Pareto optimum theories, finds out the non-domination solution in set P, finding out
The non-dominant grade of non-domination solution be set as 1, and the non-domination solution is moved into from set P in another set Q;To collection
All remaining solutions in P are closed, according to Pareto optimum theories, find out non-domination solution, and by the non-domination solution found out for the second time
Non-dominant grade is set as 2, equally moves into the non-domination solution found out for the second time in set Q from set P, and so on, until
Element in set P is 0;
3) according to crowding computational methods, the dispersibility of the non-domination solution in same non-dominant grade is calculated;
4) all non-domination solutions for obtaining non-dominant grade and dispersibility are ranked up;
5) the number z that fireworks explosion generates spark is calculated according to sequencing informationiWith fireworks explosion amplitude Ai, calculation formula is as follows:
Wherein, zmax、zmin、AmaxAnd AminIt is pre-defined constant, zmaxAnd zminIndicate respectively maximum explosive spark number and
Minimum explosive spark number, AmaxAnd AminMaximum explosion amplitude and minimum explosion amplitude are indicated respectively;I indicates fireworks after sorting
Serial number;N indicates the number of fireworks;
6) explosive spark is generated, including:Z is generated to i-th of fireworksiA explosive spark obtains the position x=of explosive spark first
xi, xiIt indicates i-th of fireworks, a node is randomly choosed from total node number, if node is selected, when selected node is
1 is changed to 0, when selected node is 0 to be changed to 1, generates explosive spark, the calculation formula of wherein a node:A=
AiRand (0,1), Ai are the explosion amplitudes of i-th of fireworks, and rand (0,1) is the random number between 0~1;
7) step 6) is repeated until generating all explosive sparks;
8) Gauss spark is generated, including:The position x of Gauss spark is obtained, x is randomly choosed from N number of fireworks, is given birth to a node
At the random number of a Gaussian Profile, if the random number in setting range, when the node is 1 to be changed to 0, when described
Node is 0 to be changed to 1, generates a Gauss spark;
9) step 8) is repeated until generating all Gauss sparks;
10) all fireworks, explosive spark and Gauss spark are ranked up according to step 2)-step 4), are obtained after sequence
Top n fireworks or explosive spark or Gauss spark are as follow-on fireworks;
11) step 2)-step 8) is repeated until reaching the iterations of setting.
2. a kind of multiple target Method for HW/SW partitioning based on fireworks algorithm according to claim 1, which is characterized in that step
It is rapid 3) described in crowding computational methods include:First to the non-domination solution in the same non-dominant grade according to an optimization mesh
Mark is ranked up, and is set any one non-dominant grade and is shared L non-domination solution, the crowding of the 1st and l-th non-domination solution is
Infinity, then the crowding of k-th of solution is using the calculating of following formula:
Crowd (k)=2 × ((t (k+1)-t (k-1))+(p (k-1)-p (k+1)))
The crowding of wherein crowd (k) k-th of solution of expression, 1<k<L, t and p are illustrated respectively in always holding for system in object space
Row time and total power consumption.
3. a kind of multiple target Method for HW/SW partitioning based on fireworks algorithm according to claim 1, which is characterized in that step
The standard of rapid 4) the described sequence is:For any two solution m and n, when the non-branch that the non-dominant grade of m-th of solution is solved than n-th
With grade is low or m-th of solution and n-th of solution are in the same non-dominant grade and n-th of the crowding ratio of m-th of solution solution
Crowding it is big, then m-th solution better than n-th solution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810164337.4A CN108446455B (en) | 2018-02-27 | 2018-02-27 | A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810164337.4A CN108446455B (en) | 2018-02-27 | 2018-02-27 | A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108446455A true CN108446455A (en) | 2018-08-24 |
CN108446455B CN108446455B (en) | 2019-09-27 |
Family
ID=63192692
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810164337.4A Active CN108446455B (en) | 2018-02-27 | 2018-02-27 | A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108446455B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109614216A (en) * | 2018-08-30 | 2019-04-12 | 广东工业大学 | A kind of cloud platform energy optimization management method based on fireworks algorithm |
CN109858617A (en) * | 2019-02-20 | 2019-06-07 | 哈尔滨工程大学 | A kind of fireworks calculation method optimizing initialization mode |
CN110097267A (en) * | 2019-04-19 | 2019-08-06 | 天津大学 | A kind of multiple target multi-robot Task Allocation based on improvement fireworks algorithm |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105550427A (en) * | 2015-12-09 | 2016-05-04 | 天津大学 | Software and hardware classifying method based on improved PBIL (Population-based Increased Learning) algorithm |
CN107395433A (en) * | 2017-08-18 | 2017-11-24 | 中南大学 | A kind of wireless sensor node dispositions method based on fireworks algorithm |
CN107658840A (en) * | 2017-09-30 | 2018-02-02 | 山东科技大学 | Distribution network failure restoration methods based on A* algorithms Yu fireworks algorithm |
-
2018
- 2018-02-27 CN CN201810164337.4A patent/CN108446455B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105550427A (en) * | 2015-12-09 | 2016-05-04 | 天津大学 | Software and hardware classifying method based on improved PBIL (Population-based Increased Learning) algorithm |
CN107395433A (en) * | 2017-08-18 | 2017-11-24 | 中南大学 | A kind of wireless sensor node dispositions method based on fireworks algorithm |
CN107658840A (en) * | 2017-09-30 | 2018-02-02 | 山东科技大学 | Distribution network failure restoration methods based on A* algorithms Yu fireworks algorithm |
Non-Patent Citations (4)
Title |
---|
XIAOHU YAN等: "An Efficient Particle Swarm Optimization for Large-Scale Hardware/Software Co-Design System", 《INTERNATIONAL JOURNAL OF COOPERATIVE INFORMATION SYSTEMS》 * |
余益科: "动态软硬件划分关键技术的研究", 《中国优秀硕士学位论文全文数据库息科技辑》 * |
韩宏业: "人工智能优化算法在软硬件划分中的应用综述", 《信息技术》 * |
韩素娟: "基于遗传和模拟退火混合的软硬件划分方法研究", 《中国优秀硕士论文全文数据库信息科技辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109614216A (en) * | 2018-08-30 | 2019-04-12 | 广东工业大学 | A kind of cloud platform energy optimization management method based on fireworks algorithm |
CN109858617A (en) * | 2019-02-20 | 2019-06-07 | 哈尔滨工程大学 | A kind of fireworks calculation method optimizing initialization mode |
CN110097267A (en) * | 2019-04-19 | 2019-08-06 | 天津大学 | A kind of multiple target multi-robot Task Allocation based on improvement fireworks algorithm |
Also Published As
Publication number | Publication date |
---|---|
CN108446455B (en) | 2019-09-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Guo et al. | Task scheduling optimization in cloud computing based on heuristic algorithm | |
CN108446455B (en) | A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm | |
CN104936186B (en) | Cognitive radio network spectrum allocation method based on cuckoo searching algorithm | |
AU2018267755A1 (en) | Array element arrangement method for L-type array antenna based on inheritance of acquired characteristics | |
CN106022601B (en) | Multi-target resource allocation method | |
CN105976122B (en) | Multi-target resource allocation system | |
CN103235743A (en) | Method for scheduling multi-target testing task based on decomposition and optimal solution following strategies | |
CN110134007A (en) | Multiple no-manned plane cooperates with target assignment method | |
Subramoney et al. | Multi-swarm PSO algorithm for static workflow scheduling in cloud-fog environments | |
Zhang et al. | A novel stochastic clustering auction for task allocation in multi-robot teams | |
Garg et al. | Multi-objective workflow grid scheduling based on discrete particle swarm optimization | |
CN108304925A (en) | A kind of pond computing device and method | |
CN104463322A (en) | Parallel and mixed artificial bee colony method of heterogeneous system | |
CN108399105B (en) | Software and hardware partitioning method based on improved brainstorming algorithm | |
CN107133877B (en) | Method for mining overlapped communities in network | |
CN111027665A (en) | Cloud manufacturing scheduling method based on improved chaotic bat swarm algorithm | |
Al-Arasi et al. | HTSCC a hybrid task scheduling algorithm in cloud computing environment | |
Li et al. | A DAG task scheduling scheme on heterogeneous computing systems using invasive weed optimization algorithm | |
CN104318035B (en) | General-to-part type triangular network multi-threading parallel generation method for massive terrain big data cloud | |
Wang et al. | A modified cuckoo search algorithm for flow shop scheduling problem with blocking | |
CN106408082A (en) | Control method and system based on region segmentation | |
CN106970840A (en) | A kind of Method for HW/SW partitioning of combination task scheduling | |
CN108415774A (en) | A kind of Method for HW/SW partitioning based on improvement fireworks algorithm | |
CN104021563B (en) | Method for segmenting noise image based on multi-objective fuzzy clustering and opposing learning | |
Reynolds et al. | Diversity as a Necessity for Sustainability in Cultural Systems: Collective Problem-Solving in Cultural Algorithms |
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 |