CN104794548A - ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method - Google Patents
ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method Download PDFInfo
- Publication number
- CN104794548A CN104794548A CN201510236495.2A CN201510236495A CN104794548A CN 104794548 A CN104794548 A CN 104794548A CN 201510236495 A CN201510236495 A CN 201510236495A CN 104794548 A CN104794548 A CN 104794548A
- Authority
- CN
- China
- Prior art keywords
- data
- arima
- counter
- sequence
- run
- 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
Landscapes
- Complex Calculations (AREA)
Abstract
The invention discloses an ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method. The method comprises the following steps: partitioning original electrical load data to form time series data according to the electrical load characteristic and strong relevance of data at the same moments every day; newly constructing a plurality of threads, synchronizing all threads, setting a counter shared globally, executing run () function to parallelize processed and partitioned data by use of the newly constructed thread, judging whether the counter in run () is locked, if so, queuing up; if not, locking the counter in the run (), and counting by adding 1; predicting by calling ARIMA according to data in the time series data, corresponding to the numerical value of the counter; and acquiring the predicted data, unlocking, repeating the steps until data in the time sequence data is completely processed. The method can remarkably improve the computing efficiency.
Description
Technical field
The present invention relates to field of computer technology, particularly relate to a kind of parallelization computing method based on the load prediction of ARIMA model.
Background technology
Power system load data presents obvious periodicity as time series, and ARIMA model (autoregression integration moving average model) has significant prediction effect to the obvious time series of periodic feature.
In prior art, usually utilize ARIMA model to carry out the analysis and prediction of Power system load data, but because this model is comparatively complicated, operand is consuming time greatly longer, therefore in engineer applied, needs improvement badly.
Summary of the invention
The object of this invention is to provide a kind of parallelization computing method based on the load prediction of ARIMA model, improve counting yield significantly.
The object of the invention is to be achieved through the following technical solutions:
Based on parallelization computing method for ARIMA model load prediction, the method comprises:
Step S11, according to Characteristics of Electric Load, every day has very strong relevance in the same time between data, raw power load data is carried out division makeup time sequence data;
Step S12, newly-built some threads, undertaken all threads synchronously, and arrange the overall situation share counter, each newly-built thread all call execution function run () come parallel processing divide after data;
Step S13, judge whether corresponding function the run () Counter that performs of current thread locks, if run () Counter locks, then waits in line; Otherwise, continue to perform step S14;
Step S14, to lock to this execution function run () Counter, counting increase by 1, and unlock;
Step S15, the data corresponding from time series data according to the numerical value of counter, and call autoregression integration moving average model ARIMA it is predicted;
Data after step S16, acquisition prediction, then jump to step S13, until the data in time series data are all disposed.
Further, call ARIMA model to carry out prediction to it and comprise:
ARIMA model is designated as ARIMA (p, q, d), and it represents the random series Y to non-stationary
tafter carrying out d difference processing, obtain stationary sequence X
t, then by stationary sequence Y
tfit to ARIMA (p, q);
Wherein, stationary sequence X
texpression formula as follows:
In above formula,
for autoregressive process, p is Autoregressive,
for autoregressive coefficient; ε
t-θ
1ε
t-1-...-θ
qε
t-qfor smoothing process, q is running mean exponent number, θ
1..., θ
qfor running mean coefficient; ε
tfor unobservable white noise sequence, form is W ~ N (0, σ
2);
As q=0, sequence X
tbecome autoregressive model AR (p), its expression formula is as follows:
As p=0, sequence X
tbecome moving average model MA (q), its expression formula is as follows:
X
t＝ε
t-θ
1ε
t-1-...-θ
qε
t-q,t∈Z。
As seen from the above technical solution provided by the invention, this programme, based on multithreading, analyzes Power system load data feature, and calls ARIMA model and carried out parallelization process, thus significantly improves counting yield.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme of the embodiment of the present invention, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawings can also be obtained according to these accompanying drawings.
The process flow diagram of a kind of parallelization computing method based on the load prediction of ARIMA model that Fig. 1 provides for the embodiment of the present invention;
What Fig. 2 provided for the embodiment of the present invention carries out based on ARIMA model the process flow diagram predicted.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on embodiments of the invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to protection scope of the present invention.
A kind of parallelization computing method based on the load prediction of ARIMA model that the embodiment of the present invention provides.As shown in Figure 1, the method mainly comprises the steps:
Step S11, according to Characteristics of Electric Load, every day has very strong relevance in the same time between data, raw power load data is carried out division makeup time sequence data.
According to the characteristic of Power system load data, current power load data can be predicted according to the raw power load data of some time before.Such as, according to before every day 12 time raw power load data predict the same day 12 time raw power load data.
This raw power load data can for it for the previous period in Power system load data, during division by not same date mutually data in the same time divide a class into, thus composition is according to the time series of moment order sequence.
Step S12, newly-built some threads, undertaken all threads synchronously, and arrange the overall situation share counter, each newly-built thread all call execution function run () come parallel processing divide after data.
In the embodiment of the present invention, thread be relation one to one with performing function run (), i.e. corresponding execution function run () of thread.
In the embodiment of the present invention, can realize based on the multithreading of JAVA, thread synchronization wherein can use synchronized (this) function to realize, counter is wherein used for recording and performs row, run () function is the method defined in java.lang.Runnable interface, be used for execution thread content, can by overriding to realize user's required function.By run () function call ARIMA (autoregression integration running mean) model in the present embodiment, its implementation is shown in Fig. 2.
Step S13, judge whether corresponding function the run () Counter that performs of current thread locks, if run () Counter locks, then waits in line; Otherwise, continue to perform step S14.
Step S14, to lock to this execution function run () Counter, counting increase by 1, and unlock.
Step S15, the data corresponding from time series data according to the numerical value of counter, and call ARIMA model it is predicted.
When counter locks and after+1, other threads then cannot operate these data, can directly unlock, and perform the operation of this step.
In the embodiment of the present invention, call ARIMA model and prediction is carried out to it comprise:
ARIMA model is designated as ARIMA (p, q, d), and it represents the random series Y to non-stationary
tafter carrying out d difference processing, obtain stationary sequence X
t, then by stationary sequence Y
tfit to ARIMA (p, q), its process as shown in Figure 2, is respectively:
1) stationarity of sequence identifies;
2) tranquilization process is carried out to non-stationary series;
3) according to the recognition rule of time series models, corresponding model is set up;
4) carry out parameter estimation, whether inspection has statistical significance;
5) carry out test of hypothesis, whether diagnosis residual sequence is white noise;
6) utilization carries out forecast analysis by the model of inspection.
Wherein, stationary sequence X
texpression formula as follows:
In above formula,
for autoregressive process, p is Autoregressive,
for autoregressive coefficient; ε
t-θ
1ε
t-1-...-θ
qε
t-qfor smoothing process, q is running mean exponent number, θ
1..., θ
qfor running mean coefficient; ε
tfor unobservable white noise sequence, its Normal Distribution (W ~ N (0, σ
2));
be parameter undetermined with θ, can obtain according to existing sequence.
As q=0, sequence X
tbecome autoregressive model AR (p), its expression formula is as follows:
As p=0, sequence X
tbecome moving average model MA (q), its expression formula is as follows:
X
t＝ε
t-θ
1ε
t-1-...-θ
qε
t-q,t∈Z。
Data after step S16, acquisition prediction, then jump to step S13, until the data in time series data are all disposed.
The such scheme of the embodiment of the present invention, based on multithreading, analyzes Power system load data feature, and calls ARIMA model and carried out parallelization process, thus significantly improves counting yield.
Through the above description of the embodiments, those skilled in the art can be well understood to above-described embodiment can by software simulating, and the mode that also can add necessary general hardware platform by software realizes.Based on such understanding, the technical scheme of above-described embodiment can embody with the form of software product, it (can be CD-ROM that this software product can be stored in a non-volatile memory medium, USB flash disk, portable hard drive etc.) in, comprise some instructions and perform method described in each embodiment of the present invention in order to make a computer equipment (can be personal computer, server, or the network equipment etc.).
The above; be only the present invention's preferably embodiment, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; the change that can expect easily or replacement, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.
Claims (2)
1., based on parallelization computing method for ARIMA model load prediction, it is characterized in that, the method comprises:
Step S11, according to Characteristics of Electric Load, every day has very strong relevance in the same time between data, raw power load data is carried out division makeup time sequence data;
Step S12, newly-built some threads, undertaken all threads synchronously, and arrange the overall situation share counter, each newly-built thread all call execution function run () come parallel processing divide after data;
Step S13, judge whether corresponding function the run () Counter that performs of current thread locks, if run () Counter locks, then waits in line; Otherwise, continue to perform step S14;
Step S14, to lock to this execution function run () Counter, counting increase by 1, and unlock;
Step S15, the data corresponding from time series data according to the numerical value of counter, and call autoregression integration moving average model ARIMA it is predicted;
Data after step S16, acquisition prediction, then jump to step S13, until the data in time series data are all disposed.
2. method according to claim 1, is characterized in that, calls ARIMA model and carries out prediction to it and comprise:
ARIMA model is designated as ARIMA (p, q, d), and it represents the random series Y to non-stationary
tafter carrying out d difference processing, obtain stationary sequence X
t, then by stationary sequence Y
tfit to ARIMA (p, q);
Wherein, stationary sequence X
texpression formula as follows:
In above formula,
for autoregressive process, p is Autoregressive,
for autoregressive coefficient; ε
t-θ
1ε
t-1-...-θ
qε
t-qfor smoothing process, q is running mean exponent number, θ
1..., θ
qfor running mean coefficient; ε
tfor unobservable white noise sequence, form is W ~ N (0, σ
2);
As q=0, sequence X
tbecome autoregressive model AR (p), its expression formula is as follows:
As p=0, sequence X
tbecome moving average model MA (q), its expression formula is as follows:
X
t＝ε
t-θ
1ε
t-1-...-θ
qε
t-q,t∈Z。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510236495.2A CN104794548A (en) | 2015-05-11 | 2015-05-11 | ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510236495.2A CN104794548A (en) | 2015-05-11 | 2015-05-11 | ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104794548A true CN104794548A (en) | 2015-07-22 |
Family
ID=53559334
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510236495.2A Pending CN104794548A (en) | 2015-05-11 | 2015-05-11 | ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104794548A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550772A (en) * | 2015-12-09 | 2016-05-04 | 中国电力科学研究院 | Online historical data tendency analysis method |
| CN105700913A (en) * | 2015-12-30 | 2016-06-22 | 广东工业大学 | Method for creation and parallel operation of lightweight bare chip codes |
| CN105787606A (en) * | 2016-03-24 | 2016-07-20 | 国网辽宁省电力有限公司电力科学研究院 | Power dispatching online trend early warning system based on ultra short term load prediction |
| CN110661253A (en) * | 2019-09-24 | 2020-01-07 | 国网(苏州)城市能源研究院有限责任公司 | Multi-type electric power elastic load combined adjusting method in building group |
| WO2021255514A1 (en) | 2020-06-15 | 2021-12-23 | Universidade Do Porto | Padding method for convolutional neural network layers adapted to perform multivariate time series analysis |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927597A (en) * | 2014-04-22 | 2014-07-16 | 国家电网公司 | Ultra-short-term wind power prediction method based on autoregression moving average model |
| CN104376388A (en) * | 2014-12-08 | 2015-02-25 | 国家电网公司 | Ultra-short period wind power prediction method based on wind speed factor control model |
-
2015
- 2015-05-11 CN CN201510236495.2A patent/CN104794548A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103927597A (en) * | 2014-04-22 | 2014-07-16 | 国家电网公司 | Ultra-short-term wind power prediction method based on autoregression moving average model |
| CN104376388A (en) * | 2014-12-08 | 2015-02-25 | 国家电网公司 | Ultra-short period wind power prediction method based on wind speed factor control model |
Non-Patent Citations (5)
| Title |
|---|
| 刘晨晖 著: "《电力系统负荷预报理论与方法》", 31 December 1987, 哈尔滨工业大学出版社 * |
| 朱陶业 等: "ARIMA模型在广西短期电力负荷预测中的应用", 《长沙电力学院学报(自然科学版)》 * |
| 袁加全 等: "基于JAVA中的多线程技术实现并行计算", 《仪器仪表用户》 * |
| 赵文硕 等: "基于并行共享挖掘算法的电力负荷预测", 《计算机与数字工程》 * |
| 郭广军 等: "基于Java多线程的并行计算技术研究及应用", 《华中师范大学学报（自然科学版）》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550772A (en) * | 2015-12-09 | 2016-05-04 | 中国电力科学研究院 | Online historical data tendency analysis method |
| CN105700913A (en) * | 2015-12-30 | 2016-06-22 | 广东工业大学 | Method for creation and parallel operation of lightweight bare chip codes |
| CN105700913B (en) * | 2015-12-30 | 2018-10-12 | 广东工业大学 | A kind of parallel operation method of lightweight bare die code |
| CN105787606A (en) * | 2016-03-24 | 2016-07-20 | 国网辽宁省电力有限公司电力科学研究院 | Power dispatching online trend early warning system based on ultra short term load prediction |
| CN110661253A (en) * | 2019-09-24 | 2020-01-07 | 国网(苏州)城市能源研究院有限责任公司 | Multi-type electric power elastic load combined adjusting method in building group |
| WO2021255514A1 (en) | 2020-06-15 | 2021-12-23 | Universidade Do Porto | Padding method for convolutional neural network layers adapted to perform multivariate time series analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104794548A (en) | ARIMA (Autoregressive integrated moving average) model load prediction based parallelization computing method | |
| Li et al. | A maximally permissive deadlock prevention policy for FMS based on Petri net siphon control and the theory of regions | |
| WO2021088207A1 (en) | Mixed deployment-based job scheduling method and apparatus for cloud computing cluster, server and storage device | |
| CN111813624B (en) | Robot execution time length estimation method based on time length analysis and related equipment thereof | |
| Leo et al. | Biodoop: bioinformatics on hadoop | |
| Zhang et al. | CPU load predictions on the computational grid | |
| CN112000457B (en) | Method, apparatus and computer program product for managing tasks in a processing system | |
| CN109936471B (en) | Multi-cluster resource allocation method and device | |
| CN105491079B (en) | The method and device of the required resource of adjustment application in cloud computing environment | |
| CN111950048A (en) | Mapping method and device for model of development view and mechanism view framework product prototype | |
| CN113849374A (en) | CPU occupancy rate prediction method, system, electronic device and storage medium | |
| CN109358595A (en) | IT operation and maintenance risk early warning method and related device | |
| CN116227647B (en) | Digital twin deduction optimization method and system for power system | |
| Liu et al. | rDenStream, a clustering algorithm over an evolving data stream | |
| CN116502761A (en) | Monitoring method and system for saving energy consumption based on big data | |
| CN107886174B (en) | Generator set maintenance arrangement method and device | |
| Eng et al. | An Estimation-Based Dynamic Load Balancing Algorithm for Efficient Load Distribution and Balancing in Heterogeneous Grid Computing Environment | |
| CN113704995B (en) | Non-dominated inequality-based unrelated parallel machine scheduling method and related equipment thereof | |
| Kramer et al. | Realizing a proactive, self-optimizing system behavior within adaptive, heterogeneous many-core architectures | |
| CN115619138A (en) | Building management and control operation method, device, equipment and medium based on digital twin | |
| Li et al. | Optimizing backup resources in the cloud | |
| CN113536042A (en) | Time series abnormity detection method, device and equipment | |
| CN111159237B (en) | System data distribution method and device, storage medium and electronic equipment | |
| CN112506622A (en) | Cloud-mobile-phone-oriented GPU computing performance prediction method and device | |
| Carastan-Santos et al. | Light-weight prediction for improving energy consumption in HPC platforms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150722 |
|
| RJ01 | Rejection of invention patent application after publication |