CN108509269A - A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision - Google Patents
A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision Download PDFInfo
- Publication number
- CN108509269A CN108509269A CN201810165373.2A CN201810165373A CN108509269A CN 108509269 A CN108509269 A CN 108509269A CN 201810165373 A CN201810165373 A CN 201810165373A CN 108509269 A CN108509269 A CN 108509269A
- Authority
- CN
- China
- Prior art keywords
- frog
- population
- position coordinates
- algorithm
- group
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- 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]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision：Initiation parameter；Initialize frog population；Iteration update starts, and calculates the concentration class factor of frog population；Judge whether the concentration class factor of frog population is more than the repopulation threshold value of setting；Carry out regeneration population；Frog is grouped；Calculate step-size in search；Optimizing in group is carried out, worst frog optimal frog jump into group in group is made；Global optimizing is carried out, worst frog in group is made to jump towards global optimum frog position；It is generated at random；All frogs are shuffled again；The position coordinates for exporting optimal frog in current population are best hardware-software partition scheme.The present invention can obtain the higher solution of quality when solving hardware-software partition problem.Solution quality when hardware-software partition problem is solved to promotion Swarm Intelligence Algorithm, is promoted hardware-software partition effect, is pushed application of the intelligence computation in terms of complex embedded system exploitation.
Description
Technical field
The present invention relates to a kind of hardware-software partition problems.The soft or hard of the algorithm that leapfrogs is shuffled based on supervision more particularly to a kind of
Part division methods.
Background technology
1, the hardware-software partition problem of the prior art
One task-set is made of many subtasks, hardware-software partition be exactly by these subtasks in software processing elements and
It is allocated on hardware processing element, different allocation plans can reach different task execution effects, therefore, hardware-software partition
Target seek to find optimal hardware-software partition scheme.As an optimization problem, task to be divided can with G come
It indicates.G=<V,E>, wherein V={ V0, V1,…Vi,…VNBe task to be divided set, task advises number of nodes N, ViIndicate the
I task node,WhereinRepresent node VkThere are software sw and hardware hw two
Kind realization method,The software for representing task node executes the time,The hardware for representing task node executes the time,It indicates to appoint
The hardware area demand of business node.E={ e1,e2,…eLIndicate existing L dependence between task node, ei=
{cia,ib|via,vib∈ V } indicate task node via,vibBetween there are dependence, vibIt will be in viaExecute later, and the two it
Between call duration time be cia,ib.It, usually will be between the execution time of all task nodes and each node in most of researchs
The sum of call duration time T (V) be used as Optimization goal, with hardware area and AsumAs constraints, then its mathematical model describes such as
Shown in following formula, wherein AreaLimit represents the size of its hardware area constraint.
min：
subject to：
For single software+mono- hardware system platform, the hardware-software partition scheme of N number of task node can use one 0,1 group
At N-dimensional vector indicate.Its 0 expression task node software realization, 1 indicates task node hardware realization.As scheme [0,1,
1,0...] software realization of task node 1, the hardware realization of task node 2, the hardware realization of task node 3, task section are indicated
The software realization ... of point 4.
2, original to shuffle application of the algorithm in hardware-software partition that leapfrog
It is a kind of novel swarm intelligence optimization to shuffle the algorithm that leapfrogs (Shuffled Frog Leap Algorithm, SFLA)
Algorithm, migration and foraging behavior of the thought source in frog group.
It is corresponded in solution space with the position coordinates of a frog in population when shuffling the algorithm that leapfrogs applied to hardware-software partition
A feasible solution, i.e., a kind of hardware-software partition scheme；Using object function T (V) as fitness function, for judging frog position
Set the quality of coordinate；The process of search optimal solution is corresponded to the process of frog-jumping transition change place coordinate；Terminated with all iteration
The position coordinates of optimal frog are exported as algorithm and are solved afterwards, that is, the optimal hardware-software partition scheme to be found.Key step is as follows：
The parameters such as frog population scale M, packet count G, algorithm end condition are determined first.Then quantity is obtained at random meets constraint for M
The solution of condition, i.e., the position of all M frogs in initialization population, subsequently into iterative process.Each iteration by grouping, more
Newly, three parts composition is shuffled.In grouping process, solution is arranged to bad by good according to the fitness value of every frog first
Sequence, the frog after sequence are assigned to successively in G group；At no point in the update process, optimizing in group is carried out first, by worst blueness in organizing
Position of the frog into group where optimal frog is once jumped, if before the position after jump is better than jump, with new position
Instead of old position, otherwise, global optimizing is carried out, the position for organizing interior worst frog optimal frog into entire population is once jumped
Jump replaces old position before if the position after jump is better than jump with new position, if still there is no more preferably position
It sets, is then generated at random, the random position for generating a new position and replacing the interior worst frog of group present；It waits in each group most
After the completion of poor frog position updates successively, frog is shuffled again, an iteration is completed.It is followed repeatedly in this manner
Ring exports the position of optimal frog in current population as best hardware-software partition when iteration reaches algorithm end condition
Scheme.
Invention content
The technical problem to be solved by the invention is to provide one kind can be promoted with Swarm Intelligence Algorithm solution software and hardware stroke
Divide solution quality when problem, and promotes the hardware-software partition side for shuffling the algorithm that leapfrogs based on supervision of hardware-software partition effect
Method.
The technical solution adopted in the present invention is：A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision, packet
Include following steps：
1) initiation parameter, including initialization frog population scale M, packet count G, hardware area constrained parameters, algorithm
Iterations；
2) frog population is initialized；
3) iteration update starts, and calculates the concentration class factor D is of frog population, the value of concentration class factor D is is bigger, explanation
The distance between frog is smaller；
4) judge whether the concentration class factor D is of frog population is more than the repopulation threshold value R_LIM of setting, be then to hold
Row step 5) otherwise then skips to step 6)；
5) regeneration population is carried out：Optimal frog position in population is retained, other frog positions in population are according to such as
Lower formula is updated：
Snew=Sori+Rand*Csize
Wherein, SoriFor the original position of frog；SnewFor the position that frog is newly-generated；Csize is the update of frog individual
Step-length, Rand are the random numbers between one 0 to 1；
6) frog is grouped：The fitness corresponding to the position coordinates of M frog is calculated, and according to fitness by well to bad pair
Frog is ranked up, and frog is divided into G group according to sorted order；
7) step-size in search is calculated：The step-size in search Step of frog is calculated according to the following formula：
Step=Stepmax-(Stepmax-Stepmin)*Dis/R_LIM
Wherein, StepmaxWith StepminRespectively maximum search step-length and minimum step-size in search；
8) optimizing in group is carried out, worst frog optimal frog jump, position coordinates after frog jump into group in group are made
If corresponding splitting scheme meets hardware area constraint and more excellent than the fitness of jump front position, the interior optimizing of frog group at
Work(replaces original position with new position coordinates, and is directly entered step 11)；Otherwise, step 9) is carried out；
9) global optimizing is carried out, worst frog in group is made to jump towards global optimum frog position, frog completes to jump
If the corresponding splitting scheme of position coordinates after jump meets hardware area constraint and, blueness more excellent than the fitness of origin-location
The frog replaces home position with the new position of frog and enters step 11) to global optimizing success；Otherwise, it enters step 10)；
10) it is generated at random, i.e., it is random to generate a new position coordinates replacement for meeting hardware area constraints
The position coordinates of worst frog in group；
11) all frogs are shuffled again, that is, are mixed the frog population after grouping and again according to fitness by well to bad
Sequence, sequence are optimal frog position first, and an iteration update is completed, and judges whether the iterations for reaching algorithm,
It is then to enter step 12), otherwise, returns to step 3) and start next round iteration；
12) it is best hardware-software partition scheme to export the position coordinates of optimal frog in current population.
Step 2) includes the task-set formed for N number of task node, firstly generates M and meets hardware area constraints
Frog position coordinates, each position coordinates be 0,1 composition N-dimensional vector.
The concentration class factor D is calculation formula of calculating frog population described in step 3) are as follows：
Wherein, x1And x2For the vector of two N-dimensionals, x1(i) and x2(i) x is indicated respectively1Vector sum x2The value of vectorial i-th dimension；
Dis_min is a threshold value；M is initialization frog population scale.
Optimal frog position described in step 5) is the fitness calculated corresponding to the position coordinates of M frog, and according to
Fitness is ranked up frog to bad by good, and sequence is optimal frog position first.
Optimizing is carried out using following formula in group described in step 8)：
S'gw=Sgw+Rand×Step×(SgB-Sgw)
Wherein, S'gwIndicate the position coordinates after worst frog jump in g groups；SgwIndicate worst frog-jumping in g groups
Position coordinates before jump；SgBIndicate the position coordinates of the optimal frog of g groups；Rand is the random number between an O to 1；Step
It is step-size in search.
Global optimizing described in step 9) is to use following formula：
S'gw=Sgw+Rand×Step×(SB-Sgw)
Wherein, S'gwIndicate the position coordinates after worst frog jump in g groups；SgwIndicate worst frog-jumping in g groups
Position coordinates before jump；Rand is the random number between one 0 to 1；SBIndicate global optimum's frog position coordinates；Step is to search
Suo Buchang.
A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision of the present invention, by introducing clumping factor to mixed
The algorithmic procedure for washing the algorithm that leapfrogs exercises supervision, and repopulation and adaptive adjustment search step are carried out respectively according to supervision result
Length carrys out the optimizing ability of boosting algorithm.Relative to the original algorithm that leapfrogs that shuffles, method of the invention is solving hardware-software partition
Quality higher solution can be obtained when problem.Solution matter when hardware-software partition problem is solved to promotion Swarm Intelligence Algorithm
Amount promotes hardware-software partition effect, pushes application of the intelligence computation in terms of complex embedded system exploitation.
Description of the drawings
Fig. 1 a are schematic diagrames before frog repopulation；
Fig. 1 b are schematic diagrames after frog repopulation.
Specific implementation mode
With reference to embodiment and attached drawing to a kind of hardware-software partition side for shuffling the algorithm that leapfrogs based on supervision of the present invention
Method is described in detail.
A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision of invention, which is characterized in that including walking as follows
Suddenly：
1) initiation parameter, including initialization frog population scale M, packet count G, hardware area constrain (AreaLimit)
Parameter, the iterations of algorithm；
2) frog population is initialized；Include the task-set formed for N number of task node, firstly generates M and meet hardware
The frog position coordinates of area-constrained condition, each position coordinates are the N-dimensional vector of 0,1 composition.
3) iteration update starts, and calculates the concentration class factor D is of frog population, the value of concentration class factor D is is bigger, explanation
The distance between frog is smaller；The concentration class factor D is calculation formula of the calculating frog population are as follows：
Wherein, x1And x2For the vector of two N-dimensionals, x1(i) and x2(i) x is indicated respectively1Vector sum x2The value of vectorial i-th dimension；
Dis_min is a threshold value；M is initialization frog population scale.
4) judge whether the concentration class factor D is of frog population is more than the repopulation threshold value R_LIM of setting, be then to hold
Row step 5) otherwise then skips to step 6)；
5) regeneration population is carried out：As illustrated in figs. 1A and ib, the optimal frog position in population is retained for regenerative process, kind
Other frog positions in group are updated according to following formula：
Snew=Sori+Rand*Csize (2)
Wherein, SoriFor the original position of frog；SnewFor the position that frog is newly-generated；Csize is the update of frog individual
Step-length, Rand are the random numbers between one 0 to 1；
The optimal frog position is the fitness calculated corresponding to the position coordinates of M frog, and according to fitness
Frog is ranked up to bad by good, sequence is optimal frog position first.
6) frog is grouped：The fitness corresponding to the position coordinates of M frog is calculated, and according to fitness by well to bad pair
Frog is ranked up, and frog is divided into G group according to sorted order；
7) step-size in search is calculated：The step-size in search Step of frog is calculated according to the following formula：
Step=Stepmax-(Stepmax-Stepmin)*Dis/R_LIM (3)
Wherein, StepmaxWith StepminRespectively maximum search step-length and minimum step-size in search；
8) optimizing in group is carried out, worst frog optimal frog jump, position coordinates after frog jump into group in group are made
If corresponding splitting scheme meets hardware area constraint and more excellent than the fitness of jump front position, the interior optimizing of frog group at
Work(replaces original position with new position coordinates, and is directly entered step 11)；Otherwise, step 9) is carried out；
Optimizing is carried out using following formula in the group：
S'gw=Sgw+Rand×Step×(SgB-Sgw) (4)
Wherein, S'gwIndicate the position coordinates after worst frog jump in g groups；SgwIndicate worst frog-jumping in g groups
Position coordinates before jump；SgBIndicate the position coordinates of the optimal frog of g groups；Rand is the random number between an O to 1；Step
It is step-size in search.
9) global optimizing is carried out, worst frog in group is made to jump towards global optimum frog position, frog completes to jump
If the corresponding splitting scheme of position coordinates after jump meets hardware area constraint and, blueness more excellent than the fitness of origin-location
The frog replaces home position with the new position of frog and enters step 11) to global optimizing success；Otherwise, it enters step 10)；
The global optimizing is to use following formula：
S'gw=Sgw+Rand×Step×(SB-Sgw) (5)
Wherein, S'gwIndicate the position coordinates after worst frog jump in g groups；SgwIndicate worst frog-jumping in g groups
Position coordinates before jump；Rand is the random number between one 0 to 1；SBIndicate global optimum's frog position coordinates；Step is to search
Suo Buchang.
10) it is generated at random, i.e., it is random to generate a new position coordinates replacement for meeting hardware area constraints
The position coordinates of worst frog in group；
11) all frogs are shuffled again, that is, are mixed the frog population after grouping and again according to fitness by well to bad
Sequence, sequence are optimal frog position first, and an iteration update is completed, and judges whether the iterations for reaching algorithm,
It is then to enter step 12), otherwise, returns to step 3) and start next round iteration；
12) it is best hardware-software partition scheme to export the position coordinates of optimal frog in current population.
For the task-set of the different scales of 50 to 1000 nodes, one kind of the invention is based on prison under identical run time
Superintend and direct that shuffle the Method for HW/SW partitioning of the algorithm that leapfrogs as shown in table 1 compared with promotion degree of the primal algorithm on solving quality.
The supervision of table 1 shuffles the algorithm that leapfrogs and shuffles the algorithm performance comparison that leapfrogs with original
Preferred example is given below：
The design parameter setting of example is as follows：
Frog population scale M=20, the iterations of packet count G=4, algorithm are 1500 iteration.It is arranged in formula (1)
Threshold value D_LIM=N/50, N is task node number, setting repopulation threshold value R_LIM=M*M*0.4=20*20*0.4=
160.The individual update step-length Csize being arranged in formula (2) is 0.1, the Step in setting formula (3)maxIt is 0.8, StepminFor
0.5.Using the execution time of all task nodes and inter-node communication time and T (V) as valuation functions, made with AreaLimit
For hardware area binding occurrence.For the task to be divided that number of nodes is N, the N-dimensional vector formed with 0 and 1 indicates the position of frog
Coordinate, i.e. a hardware-software partition scheme, supervision, which shuffles the algorithm that leapfrogs and applies to hardware-software partition, to be as follows：
(1) initiation parameter.Frog population scale M=20 is initialized, the iterations of packet count G=4, algorithm are 1500
Secondary iteration.
(2) frog population is initialized.For the task-set of N number of task node composition, firstly generates 20 and meet constraint item
The frog position coordinates of part, each position coordinates are the N-dimensional vector of 0,1 composition.
(3) iteration update starts, and calculates the concentration class factor D is of population.Shown in the calculation formula of Dis such as formula (1):
Wherein, x1And x2For the vector of two N-dimensionals, x1(i) and x2(i) x is indicated respectively1Vector sum x2The value of vectorial i-th dimension.
Dis_min is a threshold value, Dis_min=N/50.
(4) judge whether Dis more than repopulation threshold value 160 otherwise, then skips to step if so, thening follow the steps (5)
(6)。
(5) regeneration population is carried out.The position of optimal frog in population is retained, other frog positions in population according to
Formula (2) is updated
Snew=Sori+Rand*Csize (2)
Wherein, SoriFor the position of initial solution, SnewFor the position of newly-generated solution.Csize is the update step-length of individual,
Csize=0.1, Rand are the random numbers between one 0 to 1.
(6) frog is grouped.Calculate 20 frogs position coordinates corresponding to fitness value, and according to fitness by well to
It is bad that frog is ranked up, frog is divided into 4 groups according to sorted order.
(7) step-size in search is calculated.The step-size in search Step of frog is calculated according to formula (3).
Step=Stepmax-(Stepmax-Stepmin)*Dis/R_LIM (3)
Wherein, StepmaxWith StepminRespectively maximum search step-length and minimum step-size in search, Stepmax=0.8,
Stepmax=0.5.
(8) optimizing in group is carried out.Worst frog optimal frog jump into group in group.Frog position coordinates more new formula is such as
Shown in formula (4), wherein S'gwIndicate the position coordinates after worst frog jump, S in g groupsgwIndicate worst frog jump in g groups
Preceding position coordinates, SgBIndicate that the position coordinates of the optimal frog of g groups, Rand are the random numbers between one 0 to 1.
S'gw=Sgw+Rand×Step×(SgB-Sgw) (4)
If the corresponding splitting scheme of position coordinates after frog jump meets hardware area constraint and than jump front position
Fitness value it is more excellent, then optimizing success in frog group replaces original position with new position coordinates, and be directly entered step
(11)；Otherwise, step (9) is carried out.
(9) global optimizing is carried out.Worst frog jumps towards global optimum frog position in group, more new formula such as formula
(5) shown in, wherein SBIndicate global optimum's frog position coordinates：
S'gw=Sgw+Rand×Step×(SB-Sgw) (5)
If frog, which completes the corresponding splitting scheme of position coordinates after jump, meets hardware area constraint and than original
The fitness value of position is more excellent, then frog replaces original position with new position and enter step to global optimizing success
(11)；Otherwise, it enters step (10).
(10) it is generated at random.It is random to generate the new position coordinates for meeting hardware area constraints instead of group
The position coordinates of interior worst frog.
(11) all frogs are shuffled again, an iteration update complete, judge it is no reach 1500 iteration, if so, into
Enter step (12), otherwise, returns to step (3) and start next round iteration.
(12) it is best hardware-software partition scheme to export the position coordinates of optimal frog in current population.
Claims (6)
1. a kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision, which is characterized in that include the following steps：
1) initiation parameter, including initialization frog population scale M, packet count G, hardware area constrained parameters, the iteration of algorithm
Number；
2) frog population is initialized；
3) iteration update starts, and calculates the concentration class factor D is of frog population, the value of concentration class factor D is is bigger, illustrates frog
The distance between it is smaller；
4) judge whether the concentration class factor D is of frog population is more than the repopulation threshold value R_LIM of setting, be then to execute step
It is rapid 5), otherwise, then skip to step 6)；
5) regeneration population is carried out：Optimal frog position in population is retained, other frog positions in population are according to following public affairs
Formula is updated：
Snew=Sori+Rand*Csize
Wherein, SoriFor the original position of frog；SnewFor the position that frog is newly-generated；Csize is the update step-length of frog individual,
Rand is the random number between one 0 to 1；
6) frog is grouped：The fitness corresponding to the position coordinates of M frog is calculated, and according to fitness by well to bad to frog
It is ranked up, frog is divided into G group according to sorted order；
7) step-size in search is calculated：The step-size in search Step of frog is calculated according to the following formula：
Step=Stepmax-(Stepmax-Stepmin)*Dis/R_LIM
Wherein, StepmaxWith StepminRespectively maximum search step-length and minimum step-size in search；
8) optimizing in group is carried out, worst frog optimal frog jump, position coordinates after frog jump into group in group is made to correspond to
If splitting scheme meet hardware area constraint and more excellent than the fitness of jump front position, optimizing is successful in frog group, use
New position coordinates replace original position, and are directly entered step 11)；Otherwise, step 9) is carried out；
9) global optimizing is carried out, so that worst frog in group is jumped towards global optimum frog position, frog completes jump
If the corresponding splitting scheme of position coordinates afterwards meet hardware area constraint and it is more excellent than the fitness of origin-location, frog to
Global optimizing success, replaces home position with the new position of frog and enters step 11)；Otherwise, it enters step 10)；
10) it is generated at random, i.e., it is random to generate the new position coordinates for meeting hardware area constraints instead of in group
The position coordinates of worst frog；
11) all frogs are shuffled again, that is, are mixed the frog population after grouping and arranged again to bad by good according to fitness
Sequence, sequence are optimal frog position first, and an iteration update is completed, judges whether the iterations for reaching algorithm, be,
It then enters step 12), otherwise, returns to step 3) and start next round iteration；
12) it is best hardware-software partition scheme to export the position coordinates of optimal frog in current population.
2. a kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision according to claim 1, which is characterized in that
Step 2) includes the task-set formed for N number of task node, firstly generates the M frog for meeting hardware area constraints positions
Coordinate is set, each position coordinates are the N-dimensional vector of 0,1 composition.
3. a kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision according to claim 1, which is characterized in that
The concentration class factor D is calculation formula of calculating frog population described in step 3) are as follows：
Wherein, x1And x2For the vector of two N-dimensionals, x1(i) and x2(i) x is indicated respectively1Vector sum x2The value of vectorial i-th dimension；Dis_
Min is a threshold value；M is initialization frog population scale.
4. a kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision according to claim 1, which is characterized in that
Optimal frog position described in step 5), be calculate M frog position coordinates corresponding to fitness, and according to fitness by
Good to be ranked up to bad to frog, sequence is optimal frog position first.
5. a kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision according to claim 1, which is characterized in that
Optimizing is carried out using following formula in group described in step 8)：
S'gw=Sgw+Rand×Step×(SgB-Sgw)
Wherein, S'gwIndicate the position coordinates after worst frog jump in g groups；SgwIt indicates in g groups before worst frog jump
Position coordinates；SgBIndicate the position coordinates of the optimal frog of g groups；Rand is the random number between an O to 1；Step is search
Step-length.
6. a kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision according to claim 1, which is characterized in that
Global optimizing described in step 9) is to use following formula：
S'gw=Sgw+Rand×Step×(SB-Sgw)
Wherein, S'gwIndicate the position coordinates after worst frog jump in g groups；SgwIt indicates in g groups before worst frog jump
Position coordinates；Rand is the random number between one 0 to 1；SBIndicate global optimum's frog position coordinates；Step is search step
It is long.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810165373.2A CN108509269A (en) | 2018-02-27 | 2018-02-27 | A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810165373.2A CN108509269A (en) | 2018-02-27 | 2018-02-27 | A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108509269A true CN108509269A (en) | 2018-09-07 |
Family
ID=63375180
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810165373.2A Pending CN108509269A (en) | 2018-02-27 | 2018-02-27 | A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108509269A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111274682A (en) * | 2020-01-15 | 2020-06-12 | 桂林电子科技大学 | Digital microfluidic chip online test path optimization method based on frog-leaping algorithm |
| CN112598158A (en) * | 2020-12-03 | 2021-04-02 | 大连四达高技术发展有限公司 | Task scheduling method of collaborative industrial design platform based on mixed frog leaping algorithm |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573814A (en) * | 2015-01-14 | 2015-04-29 | 天津大学 | Software and hardware partition method based on multi-target shuffled frog leaps |
| CN105786759A (en) * | 2016-03-15 | 2016-07-20 | 河北工业大学 | Method for improving standard shuffled frog leaping algorithm |
| CN106980539A (en) * | 2017-03-08 | 2017-07-25 | 天津大学 | The Method for HW/SW partitioning for the algorithm that leapfrogs is shuffled based on improvement |
| CN107290793A (en) * | 2017-06-05 | 2017-10-24 | 湖南师范大学 | A kind of VHD electrical method parallel refutation method for the algorithm that leapfroged based on many strategies of weighting |
-
2018
- 2018-02-27 CN CN201810165373.2A patent/CN108509269A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573814A (en) * | 2015-01-14 | 2015-04-29 | 天津大学 | Software and hardware partition method based on multi-target shuffled frog leaps |
| CN105786759A (en) * | 2016-03-15 | 2016-07-20 | 河北工业大学 | Method for improving standard shuffled frog leaping algorithm |
| CN106980539A (en) * | 2017-03-08 | 2017-07-25 | 天津大学 | The Method for HW/SW partitioning for the algorithm that leapfrogs is shuffled based on improvement |
| CN107290793A (en) * | 2017-06-05 | 2017-10-24 | 湖南师范大学 | A kind of VHD electrical method parallel refutation method for the algorithm that leapfroged based on many strategies of weighting |
Non-Patent Citations (1)
| Title |
|---|
| 全浩军: "盲优化软硬件划分技术研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111274682A (en) * | 2020-01-15 | 2020-06-12 | 桂林电子科技大学 | Digital microfluidic chip online test path optimization method based on frog-leaping algorithm |
| CN111274682B (en) * | 2020-01-15 | 2024-01-05 | 桂林电子科技大学 | Online testing path optimization method for digital microfluidic chip based on frog-leaping algorithm |
| CN112598158A (en) * | 2020-12-03 | 2021-04-02 | 大连四达高技术发展有限公司 | Task scheduling method of collaborative industrial design platform based on mixed frog leaping algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Das et al. | Recent advances in differential evolution–an updated survey | |
| CN108053119B (en) | Improved particle swarm optimization method for solving scheduling problem of zero-waiting line shop | |
| Eiben et al. | What is an evolutionary algorithm? | |
| CN110389819B (en) | Method and system for scheduling calculation intensive batch processing tasks | |
| CN110109822B (en) | Regression testing method for carrying out test case priority ranking based on ant colony algorithm | |
| JP6724576B2 (en) | Multi-target optimization method and apparatus | |
| CN108415773B (en) | Efficient software and hardware partitioning method based on fusion algorithm | |
| CN108509269A (en) | A kind of Method for HW/SW partitioning shuffling the algorithm that leapfrogs based on supervision | |
| CN106843997B (en) | A kind of parallel virtual machine polymerization based on Spark with optimization MBBO algorithms | |
| CN111507480B (en) | Labeling method, labeling device, labeling equipment and storage medium | |
| CN104573369A (en) | Shuffled frog-leaping based division method of software and hardware | |
| CN110969362A (en) | Multi-target task scheduling method and system under cloud computing system | |
| CN105446742A (en) | Optimization method for artificial intelligence performing task | |
| CN108399105A (en) | A kind of Method for HW/SW partitioning based on improvement brainstorming algorithm | |
| CN114330715A (en) | Intelligent ammunition co-evolution task allocation method | |
| CN106980539A (en) | The Method for HW/SW partitioning for the algorithm that leapfrogs is shuffled based on improvement | |
| CN110851247B (en) | Cost optimization scheduling method for cloud workflow with constraint | |
| CN115543556A (en) | Adaptive symbolic regression method based on multitask genetic programming algorithm | |
| CN114461368A (en) | Multi-target cloud workflow scheduling method based on cooperative fruit fly algorithm | |
| CN112884368B (en) | Multi-target scheduling method and system for minimizing delivery time and delay of high-end equipment | |
| Ansótegui et al. | Boosting evolutionary algorithm configuration | |
| CN108446455A (en) | A kind of multiple target Method for HW/SW partitioning based on fireworks algorithm | |
| CN111352650A (en) | Software modularization multi-objective optimization method and system based on INSGA-II | |
| CN111026534A (en) | Workflow execution optimization method based on multi-population genetic algorithm in cloud computing environment | |
| CN110879778A (en) | Novel dynamic feedback and improved patch evaluation software automatic restoration method |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180907 |
|
| WD01 | Invention patent application deemed withdrawn after publication |