CN112163389A - Power electronic circuit optimization method based on self-adaptive distributed particle swarm optimization algorithm - Google Patents
Power electronic circuit optimization method based on self-adaptive distributed particle swarm optimization algorithm Download PDFInfo
- Publication number
- CN112163389A CN112163389A CN202010928139.8A CN202010928139A CN112163389A CN 112163389 A CN112163389 A CN 112163389A CN 202010928139 A CN202010928139 A CN 202010928139A CN 112163389 A CN112163389 A CN 112163389A
- Authority
- CN
- China
- Prior art keywords
- adaptive
- solution
- feasible
- electronic circuit
- power electronic
- 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
- 238000005457 optimization Methods 0.000 title claims abstract description 74
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 68
- 239000002245 particle Substances 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000003044 adaptive effect Effects 0.000 claims abstract description 51
- 238000004364 calculation method Methods 0.000 claims abstract description 28
- 230000005540 biological transmission Effects 0.000 claims abstract description 12
- 230000006870 function Effects 0.000 claims description 16
- 238000011161 development Methods 0.000 claims description 9
- 238000011156 evaluation Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 5
- 230000001133 acceleration Effects 0.000 claims description 3
- 230000006978 adaptation Effects 0.000 claims description 2
- 238000013461 design Methods 0.000 description 6
- 238000004088 simulation Methods 0.000 description 6
- 239000003990 capacitor Substances 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 3
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000004071 biological effect Effects 0.000 description 1
- 229960001948 caffeine Drugs 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000004100 electronic packaging Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- RYYVLZVUVIJVGH-UHFFFAOYSA-N trimethylxanthine Natural products CN1C(=O)N(C)C(=O)C2=C1N=CN2C RYYVLZVUVIJVGH-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/32—Circuit design at the digital level
- G06F30/33—Design verification, e.g. functional simulation or model checking
- G06F30/3308—Design verification, e.g. functional simulation or model checking using simulation
-
- 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)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a power electronic circuit optimization method based on a self-adaptive distributed particle swarm optimization algorithm, which comprises the following steps of: initializing parameters on a main computing node, and randomly initializing N feasible solutions according to a given element value range; placing each feasible solution to a corresponding calculation sub-node to evaluate the adaptive value of each feasible solution; judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population, and adaptively adjusting the parameters of the updating formula according to the evolution stage; after updating the parameters, updating the speed and the position of each feasible solution; performing a replacement operation; and if the ending condition of the power transmission part is reached, ending the optimization, and otherwise, returning to the adaptive value calculation of the feasible solution. The invention introduces adaptive parameters and a distributed calculation strategy, improves the diversity of algorithm population, and improves the performance of the particle swarm optimization algorithm for optimizing the power electronic circuit.
Description
Technical Field
The invention relates to the technical field of circuit optimization, in particular to a power electronic circuit optimization method based on a self-adaptive distributed particle swarm optimization algorithm.
Background
Power electronic circuits are capable of effectively controlling power transmission by adjusting supply current or voltage to adapt to a user's load, and have been widely used in various daily devices, such as mobile devices, computers, televisions, uninterruptible power supplies, and the like. With the advancement of semiconductor technology and electronic packaging technology, there is an increasing demand for automated generation of power electronic circuits.
The method for automatically designing and optimizing the circuit is mainly divided into a deterministic algorithm and a random algorithm. Deterministic algorithms, such as gradient and hill climbing, tend to fall into local optima, resulting in suboptimal combinations of elements. Moreover, some deterministic algorithms are too dependent on the choice of the initial search point and are therefore not always suitable for the optimization of power electronic circuits. In contrast, stochastic algorithms are capable of searching the solution space extensively and are therefore better suited than deterministic methods for optimizing and designing power electronic circuits. Recently, a random algorithm, i.e., an evolutionary algorithm, has attracted the attention of many researchers. Many studies on evolutionary algorithms have shown that they have great prospects for development in the design and optimization of power electronic circuits. The method is not restricted by search space restrictive assumptions, does not require assumptions such as continuity, conductibility and the like, and can find a global optimal solution with high probability from discrete, multi-extremal and noisy high-dimensional problems. Therefore, the evolutionary algorithm is well suited for the design and optimization of power electronic circuits.
The particle swarm optimization algorithm is a branch of the evolutionary algorithm and is a random search algorithm simulating the biological activities in the nature and the swarm intelligence. The particle swarm optimization algorithm has the advantages of clear definition, simplicity, practicability and the like, and is widely applied to various practical problems, such as task allocation of workflow, resource scheduling of virtual machines under cloud computing, robot control, wireless sensor networks and other fields. The particle swarm optimization algorithm is considered as a potential method in evolutionary computation, has the advantages of high convergence rate, stable solution quality and the like, and has good superiority to the optimization problem of power electronic circuit design. However, the conventional particle swarm optimization algorithm is easy to fall into a local optimal solution, resulting in early algorithm maturity.
In power electronic circuit design and optimization, different circuit elements, such as resistors, capacitors, inductors, etc., may be represented by different variables in different dimensions. However, the evaluation and simulation calculation of the conventional power electronic circuit is large in amount and long in time consumption, so that the conventional particle swarm optimization becomes not suitable for solving the circuit optimization problem.
Disclosure of Invention
In order to overcome the defects and shortcomings of the prior art, the invention provides a power electronic circuit optimization method based on a self-adaptive distributed particle swarm optimization algorithm.
The second purpose of the invention is to provide a power electronic circuit optimization system based on an adaptive distributed particle swarm optimization algorithm.
A third object of the present invention is to provide a storage medium.
It is a fourth object of the invention to provide a computing device.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides a power electronic circuit optimization method based on a self-adaptive distributed particle swarm optimization algorithm, which comprises the following steps of:
initializing algorithm parameters for optimizing a feedback network on a main computing node, and randomly initializing N feasible solutions according to a given element value range, wherein each feasible solution represents a potential power electronic circuit solution;
the feasible solution is represented as: x is the number ofi,0＝[xi,1,0,xi,2,0,...,xi,D,0]；
Wherein N represents the population size and D represents the number of elements;
placing each feasible solution to a corresponding calculation sub-node to evaluate the adaptive value of each feasible solution;
judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population, and adaptively adjusting the parameters of the updating formula according to the evolution stage;
after updating the parameters, then updating the speed and the position of each feasible solution;
and executing replacement operation: computing each newly generated solution X on a child nodeiIf the ratio is XiHistorical optimal solution pbest ofiPreferably, X is usediUpdating original pbestiOtherwise, the original pbest is reservedi；
And if the ending condition of the power transmission part is reached, ending the optimization, and otherwise, returning to the adaptive value calculation of the feasible solution.
As a preferred technical solution, the calculation function of the adaptive value is expressed as:
wherein F (x) represents the adaptive value function of the feedback network part, x represents the corresponding population individual code, vinAnd RLInput voltage and load value, v, respectivelyin_minAnd vin_maxFor minimum and maximum values of input voltage, RL_minAnd RL_maxIs the minimum and maximum value of the load, vinAnd RLStep length for changing input voltage and load respectively; f1For evaluating steady state errors at the output voltage, F2For evaluating the maximum overshoot and undershoot, and the settling time of the output voltage during start-up, F3For evaluating the steady ripple voltage on the output voltage, F4The method is used for evaluating the dynamic performance of the circuit when the input voltage and the output resistance are disturbed.
As an optimal technical scheme, judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population, and adaptively adjusting and updating parameters of a formula according to the evolution stage, wherein the method comprises the steps of estimation of the evolution stage and adaptive parameter;
the estimating step of the evolution phase comprises:
for each feasible solution, calculating Euclidean distance mean value d from the feasible solution to all other feasible solutionsiAccording to the following formula:
wherein, N and D are the size of the population scale and the dimensionality of the solution problem respectively;
the evolutionary coefficient f is calculated as follows:
wherein d isbestThe Euclidean distance mean value from the optimal solution of the population to all other particles, dminIs diMinimum value of (1), dmaxIs diMaximum value of (1);
dividing the evolution coefficient f into 4 different sets S by fuzzy logic control1、S2、S3、S4The search stage, the development stage, the convergence stage, and the jump-out stage are represented respectively.
As a preferred technical solution, the speed and the position of each feasible solution are updated by a specific calculation formula:
where ω is an inertial weight to balance global exploration and local development, c1And c2Are two acceleration factors, parameter c1Guiding the particles to learn the self optimal solution pbest so as to ensure the diversity of the whole population; parameter c2Converging the whole population to the optimal solution gbest of the current population to accelerate the convergence process of the algorithm, rand1i dAnd rand2i dIs two intervals [0,1 ]]Are uniformly distributedA random number.
As a preferred technical solution, the method further includes a distributed computing step of the master node and the slave node, and specifically includes:
in each generation, the Master node Master is responsible for the evolution operation of the whole population, and the time-consuming individual evaluation operation is carried out on the corresponding Slave node Slave; and the Master node transmits the evolved individual groups to the corresponding Slave nodes, and after the Slave nodes are evaluated, the adaptive values are returned to the Master node.
In order to achieve the second object, the present invention adopts the following technical solutions:
a power electronic circuit optimization system based on an adaptive distributed particle swarm optimization algorithm comprises the following components: the device comprises an initialization module, a feasible solution adaptive value calculation module, a self-adaptive adjustment updating module, a feasible solution speed and position updating module, a replacement module and a judgment module;
the initialization module is used for initializing algorithm parameters for optimizing a feedback network on a main computing node, and randomly initializing N feasible solutions according to a given element value range, wherein each feasible solution represents a potential power electronic circuit solution;
the feasible solution is represented as: x is the number ofi,0＝[xi,1,0,xi,2,0,...,xi,D,0]；
Wherein N represents the population size and D represents the number of elements;
the feasible solution adaptive value calculation module is used for placing each feasible solution into a corresponding calculation sub-node to evaluate the adaptive value of each feasible solution;
the self-adaptive adjustment updating module is used for judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population and self-adaptively adjusting and updating the parameters of the formula according to the evolution stage;
the speed and position updating module of the feasible solutions is used for updating the speed and position of each feasible solution;
the replacement module is to perform a replacement operation: computing each newly generated solution X on a child nodeiIf the ratio is XiHistorical optimal solution ofpbestiPreferably, X is usediUpdating original pbestiOtherwise, the original pbest is reservedi；
The judging module is used for judging whether the end condition of the power transmission part is reached, if the end condition of the power transmission part is reached, the optimization is ended, otherwise, the adaptive value calculation of the feasible solution is returned.
In order to achieve the third object, the present invention adopts the following technical solutions:
a storage medium storing a program which, when executed by a processor, implements the power electronic circuit optimization method based on an adaptive distributed particle swarm optimization algorithm as described above.
In order to achieve the fourth object, the present invention adopts the following technical means:
a computing device comprising a processor and a memory for storing processor executable programs, the processor, when executing a program stored in the memory, implementing a power electronic circuit optimization method based on an adaptive distributed particle swarm optimization algorithm as described above.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) the invention introduces adaptive parameters and a distributed calculation strategy, improves the diversity of algorithm population, and improves the performance of the particle swarm optimization algorithm for optimizing the power electronic circuit.
(2) The invention formulates the simulation of the power electronic circuit, constructs a weighted adaptive value evaluation function containing 4 targets according to the 4 types of optimization targets of the power electronic circuit, makes important contribution on the simulation formulation of the power electronic circuit and plays a key role in the optimization of the power electronic circuit.
Drawings
FIG. 1 is a basic structure diagram of a power electronic circuit of the present embodiment 1;
fig. 2 is a flowchart of optimizing a power electronic circuit by using a particle swarm optimization algorithm based on a standardized grouping strategy according to this embodiment 1;
FIG. 3 is a schematic diagram of the fuzzy logic control prediction evolution phase in this embodiment 1;
FIG. 4 is a schematic diagram of a distributed architecture framework of the embodiment 1;
fig. 5 is a schematic diagram of the buck converter of the present embodiment 1.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
As shown in fig. 1, the basic structure of a power electronic circuit includes two parts, a power transmission and a feedback network. The power transmission part comprises IPA resistance JPAn inductance and KPA capacitor; the feedback network part comprises IFA resistance JFAn inductance and KFA capacitor.
As shown in fig. 2, the present embodiment provides a power electronic circuit optimization method based on an adaptive distributed particle swarm optimization algorithm, which includes the following steps:
s1: initializing algorithm parameters for optimizing a feedback network on a master computing node, randomly initializing N feasible solutions according to a given element value range, wherein each solution (particle) Xi＝[xi1,xi2,...,xiD](1 ≦ i ≦ N) representing a potential power electronic circuit solution, where N is the population size and D is the number of elements;
s2: evaluating each particle X by placing each particle into a corresponding compute child nodeiThe fitness function is:
wherein F (x) represents the adaptive value function of the feedback network part, x represents the corresponding population individual code, vinAnd RLInput voltage and load value, v, respectivelyin_minAnd vin_maxIs an input voltageMinimum and maximum values of RL_minAnd RL_maxIs the minimum and maximum value of the load, vinAnd RLStep length for changing input voltage and load respectively; f1For evaluating steady state errors at the output voltage, F2For evaluating the maximum overshoot and undershoot, and the settling time of the output voltage during start-up, F3For evaluating the steady ripple voltage on the output voltage, F4The method is used for evaluating the dynamic performance of the circuit when the input voltage and the output resistance are disturbed. After calculation, returning the adaptive value of each particle to the main calculation node;
four objective functions F in an adaptation function F of a feedback network part1，F2，F3，F4Each defined as follows.
1.F1:
Defining a variance accumulation equation E2To evaluate v0And vrefIn NsProximity of simulation points
If E is2If the value of (A) is small, the steady state error is small, F1Will be larger. Formula F1Is defined as follows:
wherein, K1Is F1Maximum value that can be reached, K2For adjusting F1To E2The sensitivity of (2).
2.F2：
During start-up or external disturbances, a transient response v will occurdWherein
vd＝vref-v'o
F2And F4To evaluate vdIncluding 1) maximum overshoot, 2) maximum undershootAnd 3) the settling time of the response during start-up or perturbation. F2And F4Can be expressed as follows:
F2＝OV(RL,vin,X)+UV(RL,vin,X)+ST(RL,vin,X)
wherein N isTIs the number of inputs and load disturbances in the performance test.
In the above equations, OV, UV and ST are the minimum maximum overshoot, the maximum undershoot and vdAn objective function of time is established. They are defined as follows:
wherein K3Is the maximum value, M, that this objective function can reachp0Is the maximum overshoot, MpIs the actual overshoot, K4Is the pass band constant.
Wherein K5Is the maximum value, M, that this objective function can reachv0Is the maximum undershoot, MvIs a real undershoot, K6Is the pass band constant.
Wherein K7Is the maximum value, T, that this objective function can achieves0Is a constant, TsIs the actual setup time, K8For adjusting the sensitivity. T issIs defined as vdThe settling time falling within the a ± σ% passband. That is to say, the position of the nozzle is,
|vd(t)|≤0.01σ,t≥Ts
3.F3：
vothe ripple voltage on must be at the desired output vo,expNear ± Δ voWithin the limits. At F3The method for measuring X is to calculate in NsIn one simulation point, voExceeding vo,exp±ΔvoThe number of simulation points. F3Is defined as follows
Wherein, K9Is F3Maximum value that can be reached, K10Is the attenuation constant, A1Is the number of emulation points beyond the allowable sidebands. It can be seen that when A is1When increased, F3And decreases.
Particle swarm optimization algorithms require each individual to maintain two vectors, the velocity vector, in the evolution process: vi＝[vi1,vi2,...,viD]And the position vector: xi＝[xi1,xi2,...,xiD](the values of the circuit elements are stored in the positions and correspond to X in the adaptive value function), wherein i represents the number of individuals, D is the dimension of the problem, and the number of the elements to be optimized is represented in the optimized design of the power electronic circuit. The position of the particle represents the position of the corresponding solution in the solution space. In the evolution process, the particles search for a new position through evolution and updating operations, and if the new particles reach a position with a better adaptive value, the updating operation is executed to update the position of the particles to the position with the better adaptive value.
S3: judging the current evolution stage of the algorithm according to the distribution of the particles in the current population, and adaptively adjusting the parameters of the updated formula according to the evolution stage, wherein the method comprises the steps of estimation of the evolution stage and adaptive parameter, and specifically comprises the following steps:
estimation of the evolution phase:
step 1: for each particle Xi(1. ltoreq. i. ltoreq.N), calculating the particleMean Euclidean distance d to all other particlesiAccording to the following formula:
wherein, N and D are the size of the population scale and the dimension of the solution problem (i.e. the number of elements to be optimized), respectively;
step 2: the evolutionary coefficient f is calculated as follows:
wherein d isbestThe Euclidean distance mean value from the optimal solution of the population to all other particles, dminIs diMinimum value of (1), dmaxIs diMaximum value of (2). It is clear that f is [0,1 ]]；
And step 3: as shown in FIG. 3, f is divided into 4 different sets S using fuzzy logic control1、S2、S3、S4Respectively representing an exploration stage, a development stage, a convergence stage and a jump-out stage;
it is noted that when f belongs to two stages, for example, f is 0.28, when f belongs to S2、S3The current evolution stage is estimated according to the state of the previous stage; if the previous stage is S2I.e. the development phase, then the current phase is still divided into S according to the state stability2(ii) a Similarly, if the previous stage is S3The current stage is still divided into S3(ii) a And if the previous stage is S1Or S4Judging the current stage according to the sequence of the evolution stage; it is obvious that the evolutionary sequence of the particle swarm algorithm can be represented as S1→S2→S3→S4→S1→ …, repeating; therefore, when the previous stage is S1Or S4When the current stage is divided into S2；
Parameter adaptive operation:
and (3) an exploration phase: the population diversity should be maintained at this stage in order to explore other places in the search space, so the particles should be made to learn their historical optimal solutions pbest as much as possible to maintain their own performance, rather than learning the optimal solution gbest of the current population, since it is likely to be only a local optimal solution; therefore, c should be increased at this stage1While decreasing c2；
And (3) in a development stage: this stage should enhance the local search capability of the particle, thus increasing c appropriately1To maintain a large value to guarantee local search of individuals; meanwhile, c is properly reduced because the population optimal solution gbest is not necessarily the global optimal solution2Maintaining a small value to prevent excessive convergence and avoid falling into a local optimal value;
and (3) a convergence stage: at this stage, the whole population may have found the optimal solution, so in order to accelerate the convergence speed of the whole population, the particles in the population should learn the optimal solution gbest of the current population as much as possible, so c is added2The impact of gbest can be improved; however, in order to prevent excessive convergence while also securing population diversity, the ability to locally search each individual is increased, so c should be appropriately increased2While appropriately increasing c1；
And (3) jumping out stage: the optimal solution of the population in the stage jumps out of the whole population, which shows that the current population falls into local optimality, and all individuals should jump out of the current local optimality immediately following the gbest; so c should be increased2While decreasing c1。
S4: after updating the parameters, the velocity and position of each particle are then updated, as follows:
where ω is the inertiaSex weights to balance global exploration and local development, c1And c2Is two acceleration coefficients, where the parameter c1Guiding the particles to learn the self optimal solution pbest so as to ensure the diversity of the whole population; parameter c2Converging the whole population to the optimal solution gbest of the current population to accelerate the convergence process of the algorithm, rand1i dAnd rand2i dIs two intervals [0,1 ]]Uniformly distributed random numbers of (a);
s5: and executing replacement operation: computing each newly generated solution X on a child nodeiIf the ratio is XiHistorical optimal solution pbest ofiPreferably, X is usediUpdating original pbestiOtherwise, the original pbest is reservediFor a maximization problem, the update formula is as follows:
at the same time, if XiIs better than the optimal solution gbest of the whole population, and X is usediAnd updating the original gbest, otherwise, keeping the original gbest, and updating the formula as follows:
fit is an adaptive value evaluation function;
s6: if the end condition of the power transfer portion is reached, i.e. the algorithm has reached the maximum number of iterations, the optimization procedure is ended, otherwise it returns to step S2.
As shown in fig. 4, the present embodiment also adopts a distributed policy of a master-slave node model, where a master node is responsible for the evolution operation of the whole population, and a time-consuming circuit evaluation operation is implemented in a distributed manner by a plurality of sub-computing nodes, so as to save the algorithm running time.
Aiming at the problem of time consumption for evaluating optimization of the power electronic circuit, the embodiment adopts a distributed framework of a master node and a slave node, so that the calculation time is effectively shortened; specifically, theIn each generation, the Master node Master is responsible for the evolution of the entire population, and the time-consuming individual evaluation operation (i.e., the step S2 of evaluating the individual fitness value, evaluating each target F1-F4And the accumulated target F) is put on the corresponding Slave node Slave; the Master node transmits the evolved individual groups to corresponding Slave nodes, and after the Slave nodes are evaluated, the adaptive values are returned to the Master node; by utilizing distributed resources of a plurality of nodes, the evaluation running time is effectively shortened.
As shown in fig. 5, a test is performed by taking an optimized design of a buck converter as an example, and the graph includes 8 resistors, 4 capacitors and an inductor, so that the dimension of the problem is 13 dimensions, the value range of each dimension is determined according to the value range of a corresponding component, each particle in the particle swarm optimization represents a potential solution, and the value of each dimension in each solution represents the parameter of the corresponding component. Through algorithm optimization, parameter values of the optimized circuit components are found, and the method is further proved to be very effective.
In the embodiment, the evolutionary phase is judged by utilizing the distribution of particles in the population, and the algorithm parameters are adaptively adjusted according to the evolutionary phase. Meanwhile, distributed calculation of a master node model and a slave node model is introduced into the particle swarm optimization algorithm, so that the program running time can be obviously shortened, and the performance of the particle swarm optimization algorithm for optimizing the power electronic circuit is improved.
Example 2
The embodiment provides a power electronic circuit optimization system based on a self-adaptive distributed particle swarm optimization algorithm, which includes: the device comprises an initialization module, a feasible solution adaptive value calculation module, a self-adaptive adjustment updating module, a feasible solution speed and position updating module, a replacement module and a judgment module;
the initialization module is used for initializing algorithm parameters for optimizing the feedback network on the main computing node, and randomly initializing N feasible solutions according to a given element value range, wherein each feasible solution represents a potential power electronic circuit solution;
the feasible solution is represented as: x is the number ofi,0＝[xi,1,0,xi,2,0,...,xi,D,0]；
Wherein N represents the population size and D represents the number of elements;
the feasible solution adaptive value calculating module is used for placing each feasible solution to a corresponding calculating sub-node to evaluate the adaptive value of each feasible solution;
the self-adaptive adjustment updating module is used for judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population and self-adaptively adjusting the parameters of the updating formula according to the evolution stage;
the feasible solution speed and position updating module is used for updating the speed and position of each feasible solution;
the replacement module is used for executing replacement operation: computing each newly generated solution X on a child nodeiIf the ratio is XiHistorical optimal solution pbest ofiPreferably, X is usediUpdating original pbestiOtherwise, the original pbest is reservedi；
The judging module is used for judging whether the ending condition of the power transmission part is reached, if the ending condition of the power transmission part is reached, the optimization is ended, otherwise, the adaptive value calculation of the feasible solution is returned.
Example 3
The present embodiment further provides a storage medium, which may be a storage medium such as a ROM, a RAM, a magnetic disk, an optical disk, or the like, where one or more programs are stored, and when the programs are executed by a processor, the power electronic circuit optimization method based on the adaptive distributed particle swarm optimization algorithm in embodiment 1 is implemented.
Example 4
The embodiment also provides a computing device, where the computing device may be a desktop computer, a notebook computer, a smart phone, a PDA handheld terminal, a tablet computer, or other terminal devices with a display function, the computing device includes a processor and a memory, the memory stores one or more programs, and when the processor executes the programs stored in the memory, the power electronic circuit optimization method based on the adaptive distributed particle swarm optimization algorithm in embodiment 1 is implemented.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.
Claims (8)
1. A power electronic circuit optimization method based on a self-adaptive distributed particle swarm optimization algorithm is characterized by comprising the following steps:
initializing algorithm parameters for optimizing a feedback network on a main computing node, and randomly initializing N feasible solutions according to a given element value range, wherein each feasible solution represents a potential power electronic circuit solution;
the feasible solution is represented as: x is the number ofi,0＝[xi,1,0,xi,2,0,...,xi,D,0]；
Wherein N represents the population size and D represents the number of elements;
placing each feasible solution to a corresponding calculation sub-node to evaluate the adaptive value of each feasible solution;
judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population, and adaptively adjusting the parameters of the updating formula according to the evolution stage;
after updating the parameters, then updating the speed and the position of each feasible solution;
and executing replacement operation: computing each newly generated solution X on a child nodeiIf the ratio is XiHistorical optimal solution pbest ofiPreferably, X is usediUpdating original pbestiOtherwise, the original pbest is reservedi；
And if the ending condition of the power transmission part is reached, ending the optimization, and otherwise, returning to the adaptive value calculation of the feasible solution.
2. The power electronic circuit optimization method based on the adaptive distributed particle swarm optimization algorithm according to claim 1, wherein the calculation function of the adaptation value is expressed as:
wherein F (x) represents the adaptive value function of the feedback network part, x represents the corresponding population individual code, vinAnd RLInput voltage and load value, v, respectivelyin_minAnd vin_maxFor minimum and maximum values of input voltage, RL_minAnd RL_maxIs the minimum and maximum value of the load, vinAnd RLStep length for changing input voltage and load respectively; f1For evaluating steady state errors at the output voltage, F2For evaluating the maximum overshoot and undershoot, and the settling time of the output voltage during start-up, F3For evaluating the steady ripple voltage on the output voltage, F4The method is used for evaluating the dynamic performance of the circuit when the input voltage and the output resistance are disturbed.
3. The power electronic circuit optimization method based on the adaptive distributed particle swarm optimization algorithm according to claim 1, wherein the current evolution stage of the algorithm is judged according to the distribution of feasible solutions in the current population, and parameters of an update formula are adaptively adjusted according to the evolution stage, including the steps of estimation of the evolution stage and adaptive parameter;
the estimating step of the evolution phase comprises:
for each feasible solution, calculating Euclidean distance mean value d from the feasible solution to all other feasible solutionsiAccording to the following formula:
wherein, N and D are the size of the population scale and the dimensionality of the solution problem respectively;
the evolutionary coefficient f is calculated as follows:
wherein d isbestThe Euclidean distance mean value from the optimal solution of the population to all other particles, dminIs diMinimum value of (1), dmaxIs diMaximum value of (1);
dividing the evolution coefficient f into 4 different sets S by fuzzy logic control1、S2、S3、S4The search stage, the development stage, the convergence stage, and the jump-out stage are represented respectively.
4. The power electronic circuit optimization method based on the adaptive distributed particle swarm optimization algorithm according to claim 1, wherein the speed and the position of each feasible solution are updated according to a specific calculation formula:
where ω is an inertial weight to balance global exploration and local development, c1And c2Are two acceleration factors, parameter c1Guiding the particles to learn the self optimal solution pbest so as to ensure the diversity of the whole population; parameter c2Converging the whole population to the optimal solution gbest of the current population to accelerate the convergence process of the algorithm, rand1i dAnd rand2i dIs two intervals [0,1 ]]Uniformly distributed random numbers.
5. The power electronic circuit optimization method based on the adaptive distributed particle swarm optimization algorithm according to claim 1, further comprising a distributed calculation step of master and slave nodes, specifically comprising:
in each generation, the Master node Master is responsible for the evolution operation of the whole population, and the time-consuming individual evaluation operation is carried out on the corresponding Slave node Slave; and the Master node transmits the evolved individual groups to the corresponding Slave nodes, and after the Slave nodes are evaluated, the adaptive values are returned to the Master node.
6. A power electronic circuit optimization system based on an adaptive distributed particle swarm optimization algorithm is characterized by comprising the following steps: the device comprises an initialization module, a feasible solution adaptive value calculation module, a self-adaptive adjustment updating module, a feasible solution speed and position updating module, a replacement module and a judgment module;
the initialization module is used for initializing algorithm parameters for optimizing a feedback network on a main computing node, and randomly initializing N feasible solutions according to a given element value range, wherein each feasible solution represents a potential power electronic circuit solution;
the feasible solution is represented as: x is the number ofi,0＝[xi,1,0,xi,2,0,...,xi,D,0]；
Wherein N represents the population size and D represents the number of elements;
the feasible solution adaptive value calculation module is used for placing each feasible solution into a corresponding calculation sub-node to evaluate the adaptive value of each feasible solution;
the self-adaptive adjustment updating module is used for judging the current evolution stage of the algorithm according to the distribution of feasible solutions in the current population and self-adaptively adjusting and updating the parameters of the formula according to the evolution stage;
the speed and position updating module of the feasible solutions is used for updating the speed and position of each feasible solution;
the replacement module is to perform a replacement operation: computing each newly generated solution X on a child nodeiIf the ratio is XiHistorical optimal solution pbest ofiPreferably, X is usediUpdating original pbestiOtherwise, the original pbest is reservedi；
The judging module is used for judging whether the end condition of the power transmission part is reached, if the end condition of the power transmission part is reached, the optimization is ended, otherwise, the adaptive value calculation of the feasible solution is returned.
7. A storage medium storing a program, characterized in that the program, when executed by a processor, implements the power electronic circuit optimization method based on adaptive distributed particle swarm optimization algorithm according to any of claims 1-5.
8. A computing device comprising a processor and a memory for storing processor executable programs, characterized in that the processor, when executing a program stored by the memory, implements the method for power electronic circuit optimization based on adaptive distributed particle swarm optimization algorithm according to any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010928139.8A CN112163389A (en) | 2020-09-07 | 2020-09-07 | Power electronic circuit optimization method based on self-adaptive distributed particle swarm optimization algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010928139.8A CN112163389A (en) | 2020-09-07 | 2020-09-07 | Power electronic circuit optimization method based on self-adaptive distributed particle swarm optimization algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112163389A true CN112163389A (en) | 2021-01-01 |
Family
ID=73857883
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010928139.8A Pending CN112163389A (en) | 2020-09-07 | 2020-09-07 | Power electronic circuit optimization method based on self-adaptive distributed particle swarm optimization algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112163389A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268869A (en) * | 2014-09-23 | 2015-01-07 | 中山大学 | Automatic multilevel-resolution remote-sensing image registration method based on particle swarm optimization |
| CN105740977A (en) * | 2016-01-28 | 2016-07-06 | 福州大学 | Multi-target particle swarm-based power outage management optimization method |
| CN108053119A (en) * | 2017-12-15 | 2018-05-18 | 兰州理工大学 | A kind of Modified particle swarm optimization algorithm for solving zero-waiting Flow Shop Scheduling |
| CN108108532A (en) * | 2017-12-06 | 2018-06-01 | 华南理工大学 | With the method for particle cluster algorithm optimization power electronic circuit |
| CN108573300A (en) * | 2018-04-18 | 2018-09-25 | 辽东学院 | Particle cluster algorithm feedback adaptive rate parameter optimization method and full-order state observer |
| CN109670210A (en) * | 2018-11-26 | 2019-04-23 | 华南理工大学 | A kind of method of optimization for power electronic circuit based on parallel distributed particle swarm algorithm |
-
2020
- 2020-09-07 CN CN202010928139.8A patent/CN112163389A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268869A (en) * | 2014-09-23 | 2015-01-07 | 中山大学 | Automatic multilevel-resolution remote-sensing image registration method based on particle swarm optimization |
| CN105740977A (en) * | 2016-01-28 | 2016-07-06 | 福州大学 | Multi-target particle swarm-based power outage management optimization method |
| CN108108532A (en) * | 2017-12-06 | 2018-06-01 | 华南理工大学 | With the method for particle cluster algorithm optimization power electronic circuit |
| CN108053119A (en) * | 2017-12-15 | 2018-05-18 | 兰州理工大学 | A kind of Modified particle swarm optimization algorithm for solving zero-waiting Flow Shop Scheduling |
| CN108573300A (en) * | 2018-04-18 | 2018-09-25 | 辽东学院 | Particle cluster algorithm feedback adaptive rate parameter optimization method and full-order state observer |
| CN109670210A (en) * | 2018-11-26 | 2019-04-23 | 华南理工大学 | A kind of method of optimization for power electronic circuit based on parallel distributed particle swarm algorithm |
Non-Patent Citations (2)
| Title |
|---|
| ZHI-HUI ZHAN等: "Adaptive Particle Swarm Optimization", 《IEEE TRANS SYST MAN CYBERN B CYBERN》, vol. 39, no. 6, 31 December 2009 (2009-12-31), pages 1362 - 1381 * |
| 唐剑等: "动态环境下分布式自适应粒子群优化算法", 《系统仿真学报》, vol. 21, no. 17, 30 September 2009 (2009-09-30), pages 5431 - 5435 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2019109757A1 (en) | Method for using particle swarm algorithm to optimize power electronic circuit | |
| CN106600059B (en) | Intelligent power grid short-term load prediction method based on improved RBF neural network | |
| Panda et al. | Comparison of particle swarm optimization and genetic algorithm for FACTS-based controller design | |
| Sivalingam et al. | A modified whale optimization algorithm-based adaptive fuzzy logic PID controller for load frequency control of autonomous power generation systems | |
| CN112700060B (en) | Station terminal load prediction method and prediction device | |
| CN111459022B (en) | Device parameter adjustment method, device control apparatus, and computer-readable storage medium | |
| CN115293052A (en) | Power system active power flow online optimization control method, storage medium and device | |
| CN115833101B (en) | Power scheduling method, device, electronic equipment and storage medium | |
| CN101446997A (en) | Power electronic circuit optimization method based on particle swarm algorithm | |
| CN116914751A (en) | Intelligent power distribution control system | |
| CN117670586A (en) | Power grid node carbon factor prediction method and system based on graph neural network | |
| Fu et al. | The distributed economic dispatch of smart grid based on deep reinforcement learning | |
| CN114884063A (en) | Distributed power supply and energy storage voltage control method and device considering interval coordination | |
| CN112163389A (en) | Power electronic circuit optimization method based on self-adaptive distributed particle swarm optimization algorithm | |
| CN116826762A (en) | Intelligent power distribution network voltage safety control method, device, equipment and medium thereof | |
| CN114942799B (en) | Workflow scheduling method based on reinforcement learning in cloud edge environment | |
| CN109670210A (en) | A kind of method of optimization for power electronic circuit based on parallel distributed particle swarm algorithm | |
| Jasmin et al. | Reinforcement learning solution for unit commitment problem through pursuit method | |
| CN112163387B (en) | Power electronic circuit optimization method based on brain storm algorithm and application thereof | |
| CN112163388A (en) | Power electronic circuit optimization method based on self-adaptive distributed differential evolution algorithm | |
| Nishat et al. | Applying Particle Swarm Optimization for Investigating the Stability of DC-DC Buck-Boost Converter: A Software Based Approach | |
| Wang et al. | Automatic Voltage Control of Differential Power Grids Based on Transfer Learning and Deep Reinforcement Learning | |
| Qin et al. | Data-based reinforcement learning with application to wind turbine pitch control | |
| Fagan et al. | Dynamic multi-agent reinforcement learning for control optimization | |
| CN114697974B (en) | Network coverage optimization method and device, electronic equipment and storage medium |
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 |