Research Article  Open Access
Zhilong Wang, Feng Liu, Jie Wu, Jianzhou Wang, "A Hybrid Forecasting Model Based on Bivariate Division and a Backpropagation Artificial Neural Network Optimized by Chaos Particle Swarm Optimization for DayAhead Electricity Price", Abstract and Applied Analysis, vol. 2014, Article ID 249208, 31 pages, 2014. https://doi.org/10.1155/2014/249208
A Hybrid Forecasting Model Based on Bivariate Division and a Backpropagation Artificial Neural Network Optimized by Chaos Particle Swarm Optimization for DayAhead Electricity Price
Abstract
In the electricity market, the electricity price plays an inevitable role. Nevertheless, accurate price forecasting, a vital factor affecting both government regulatory agencies and public power companies, remains a huge challenge and a critical problem. Determining how to address the accurate forecasting problem becomes an even more significant task in an era in which electricity is increasingly important. Based on the chaos particle swarm optimization (CPSO), the backpropagation artificial neural network (BPANN), and the idea of bivariate division, this paper proposes a bivariate division BPANN (BDBPANN) method and the CPSOBDBPANN method for forecasting electricity price. The former method creatively transforms the electricity demand and price to be a new variable, named DV, which is calculated using the division principle, to forecast the dayahead electricity by multiplying the forecasted values of the DVs and forecasted values of the demand. Next, to improve the accuracy of BDBPANN, chaos particle swarm optimization and BDBPANN are synthesized to form a novel model, CPSOBDBPANN. In this study, CPSO is utilized to optimize the initial parameters of BDBPANN to make its output more stable than the original model. Finally, two forecasting strategies are proposed regarding different situations.
1. Introduction
The price of electricity is not only related to the interests of the market participants but also affected by many aspects of the relevant society and the economy. Thus, the price of electricity has been the subject of considerable research. Accurate forecasting of electricity price is of great significance, both for government regulatory agencies and for public power companies. However, the electricity prices exhibit large fluctuations, and the characteristics of the electricity prices in various electricity markets are notably different. Therefore, it is notably difficult to predict all of the electricity prices using only one model. As indicated by Anbazhagan and Kumarappan, in general, the forecasting errors for the electricity price vary from approximately 5% to 36%, which are relatively high compared with the forecasting errors for the electricity load (usually in the range from 1% to 3%). Thus, there is still an urgent demand for exploiting new electricity price forecasting models to improve the electricity price forecasting accuracy [1].
Existing forecasting models for the electricity price can be categorized by different classification criteria. Two criteria are adopted frequently, that is, forecasting models based classification and forecasting horizons based classification. According to the former classification criterion, the electricity price forecasting models can be generally classified into three categories [2]: (i) the stochastic models, (ii) the causal forecasting approaches, and (iii) the artificial intelligence based methods. According to the latter criterion, forecasting models for electricity price consist of three types: the shortterm, the midterm, and the longterm models. However, as is known, most forecasting models cannot be sorted into a single category by the forecasting models based classification rule because increasing numbers of hybrid or combined models are being developed to obtain electricity price results with superior performance. Therefore, this paper will survey and summarize the previous electricity price forecasting models according to their forecasting horizons. The shortterm electricity price forecasting is usually known as the dayahead electricity price forecasting. During the recent few decades, a large number of dayahead electricity price forecasting models have been exploited for their importance in establishing bidding strategies for the spot market [3]. A twostage model derived from panel cointegration and a particle filter was proposed by Li et al. [4] to forecast the dayahead electricity data of Pennsylvania, New Jersey, Maryland (PJM). Zhang et al. [5] presented the performance of a hybrid approach composed of the wavelet transform (WT) method, the autoregressive integrated moving average (ARIMA) model, the least squares support vector machine (LSSVM), and the particle swarm optimization in forecasting the dayahead electricity price of New South Wales, Australia. The WT and ARIMA algorithms were also used by Tan et al. [6] to build a hybrid model with the generalized autoregressive conditional heteroskedasticity to examine the dayahead electricity price forecasting performance of PJM and Spain. A dayahead price forecasting model, which was constructed by a feature selection approach and a cascaded neural network algorithm, was developed by Amjady and Keynia [7] to perform an electricity price survey in Spain and Australia. Several other dayahead electricity price forecasting models can be found in [8–15]. Accurate midterm electricity price forecasting is essential in assisting the market players to define their contracts and hedging policies. Horizons for midterm electricity price forecasting usually last from one month ahead to six months ahead. Yan and Chowdhury [16] presented a hybrid forecasting model of the midterm electricity clearing price on the basis of the LSSVM and autoregressive moving average with external input models. A multiple support vector machine approach was also applied by the same two authors [17] to the electricity market clearing price forecasting by using the PJM interconnection data. Compared to the shortterm and midterm electricity price forecasting tasks, the longterm electricity price forecasting is conducted on a much wider horizon to provide the market players with the ability to make expansion plans. Pao [3] adopted an artificial neural network model to survey the longterm electricity price forecasting results on the European Energy Exchange market. However, studies regarding the longterm electricity price forecasting are relatively few compared to the number of studies on shortterm and midterm electricity price forecasting due to the relatively long forecasting horizons required.
The accuracy of electricity price forecasting is affected by many factors, such as historical price, temperature, and load, among which the effect of load on electricity price forecasting is quite significant. Much of the relevant literature attempted to improve the accuracy of the electricity price by considering the impact of electricity load on the electricity price. For example, Lin et al. [18] regarded the load as one of the inputs to the enhanced probability neural network proposed in their paper to demonstrate its performance in the electricity price forecasting. Load was also contained in the input variables of the enhanced radial basis function network proposed by Lin et al. [15]. Singhal and Swarup [19] also used the historical load and forecasted load as inputs to the artificial neural network models to obtain the electricity price forecasting results. Although the specific models used in these studies are different, these studies have one factor in common: the impact of the load on the electricity price in these studies was embodied by directly using the load as one of the input variables and the electricity price as the output variable, rather than to transform both electricity demand and price into a new variable. Based on this idea of transformation, a new hybrid electricity model is proposed in this paper.
2. Backpropagation Artificial Neuron Network (BPANN)
An artificial neural network is widely used to validate the underlying relationship between the relevant dependent and independent variables. The advantage of the use of neural networks compared to the traditional methods lies in their capacity to analyze complicated patterns quickly with a low error. In addition, in neural networks, no assumptions are required regarding the characteristics of the underlying distribution of the data [20]. Among all of the neural networks, the backpropagation artificial neural network (BPANN) is the most frequently used neural network. The basic BPANN consists of three parts: the input layer part, the hidden layer part (or middle layer part, as one or more layers can be contained in this part), and the output layer part. Generally, the number of neurons in the input layer and in the output layer is determined according to the desired input variables and output variables, respectively, while there is no fixed rule to choose the number of the neurons in the hidden layer part. For a BPANN with neurons in the current layer, the output of the th neuron in the adjacent next layer is expressed as where is the weight connection from the neuron of the current layer to the neuron of the adjacent next layer, is the threshold of the th neuron, and is the activation function, which usually consists of three types [21]:(i) is the linear function defined as (ii) is the logistic sigmoid (logsig) function defined as (iii) is the hyperbolic sigmoid (tansig) function defined as
To validate the performance of a BP neural network, the sample data are usually divided into three parts: the training data, the validation data, and the testing data. If there are inputoutput pairs in the training data in total and the number of the neurons in the output layer is , then the objective function of the BPANN is given by where and are the desired and actual outputs, respectively, of the th neuron obtained by the th inputoutput pair.
In the operation process of the BPANN, the weight is updated according to the gradient descent rules expressed as follows [22]: where is the learning rate parameter. In this paper, BPANN has three layers, the weights and thresholds of which are denoted by , and , . The activation function of first layer is a logistic sigmoid function, and that of the second layer is a linear function.
3. Bivariate Division BPANN (BDBPANN)
In Section 1, it is mentioned that electricity price and demand have a closed relationship. In addition, Singhal and Swarup obtained the result that the price forecasting values using the neural network model indicate that the electricity price in the deregulated markets is strongly dependent on the trend in the load demand and the clearing price [19]. Therefore, to promote the accuracy of price forecasting, a bivariate division model based on BPANN, the price databases, and the demand databases is proposed. This hybrid approach skillfully transforms both of the two variables, demand and price, into one, as denoted by DV, using the dividing principle. This principle can be shown as Next, a sample matrix of DV is used to obtain the parameters of BPANN, which will provide the forecasting values of DV. Meanwhile, the network of electricity demand is also obtained by training its BPANN. Eventually, the electricity price forecasting values can be calculated using the following simple formula: Details of BDBPANN can be shown by these steps.
Step 1. Obtain the values of DV by using .
Step 2. Perform preprocessing of the training set of electricity demand using the following formula: In addition, the testing set will be mapped by Note that the inputs and the outputs of the training sets and testing sets themselves are column vectors.
Step 3. Use to train the BPANN and obtain BPANN_{Demand}.
Step 4. Obtain the mapping values of the inputs of through
Step 5. Obtain the forecasting values by postprocessing of the output by
Step 6. Replace the electricity demand with DV values and repeat Steps 2–5 to obtain the forecasting values of the DVs.
Step 7. Obtain the forecasting price by
Figure 1 shows the specific process of BDBPANN.
4. Chaos Particle Swarm Optimization BDBPANN (CPSOBDBPANN)
The particle swarm optimization (PSO) algorithm was first developed by Kennedy and Eberhart [23] as a populationbased optimization approach. In a PSO system, there are multiple candidate solutions, and each solution is named a “particle.” Each particle is captured by two factors: the velocity and the position.
4.1. Chaos Particle Swarm Optimization (CPSO)
In a standard PSO, at iteration , for the th particle, its velocity () and the position () are updated by [24] where is the inertia weight, rand1 and rand2 are two independent random variables drawn from a uniform distribution with range , and and are positive constant parameters that are known as the learning parameters. The remaining two variables and denote the personal best position of the th particle in history and the current global best position, respectively; that is, by substituting these two values into the predetermined objective function, the optimal values of the th particle and the whole particles can be obtained.
Although the standard PSO is widely applied in the parameter optimization field, it had disadvantages of easily remaining in the local optimization of the parameters, and its convergence speed is slow. As indicated by Singhal and Swarup [19], the use of the chaos will accelerate the convergence speed due to the nonrepetitive nature of chaos. Thus, the chaotic PSO (CPSO) algorithm is adopted in this paper to obtain the optimal parameter values with a higher convergence speed. The improvement of the CPSO algorithm over the standard PSO is that it initializes the position of the particles by the following formula: where is the control variable located in the range .
4.2. CPSOBDBPANN
In the classical training process of an artificial neuron network, especially for that of BPANN, the initial parameters, and along with and , often, are random matrices or vectors and consequently easily fall into the local optimal solution [25]. Chaos particle swarm optimization (abbreviated as CPSO) is an effective artificial algorithm that has the ability to achieve global optimization. CPSO is based on the basic concept of particle swarm optimization and the theory of chaos. In this paper, CPSO is used to achieve the best initial parameters of the BPANN. The improved network outperforms the previous one and has an improved ability to determine the global optimal parameters, with a lower probability of overfitting and a higher forecasting accuracy. Main steps of this CPSOBPANN are described as follows.
Step 1. Reshape the parameters , and, of the BPANN to one vector and denote the length of the vector by .
Step 2. Perform the initialization of the CPSO’s swarm, whose length is .
Step 3. Choose the fitness evaluation of CPSO to be the mean squared error (MSE):
Step 4. Execute the optimizing process of the CPSO and obtain the best swarm.
Step 5. Reshape the best swarm into four parameters of the network, and then reconstruct the BPANN by using these optimal matrices and vectors. The resulting network is called CPSOBPANN.
Next, the BDBPANN will replace the BPANN’s part of the CPSOBPANN. The process of the new forecasting model is as follows.
Step 1. Obtain the values of DV by using .
Step 2. Perform the preprocessing of the training set of the electricity demand using the following formula: And the testing set will be mapped by Note that the inputs and outputs of the training sets and the testing sets are column vectors.
Step 3. Construct the BPANN based on the structure of the data.
Step 4. Reshape the parameters, , and , , of the BPANN into one vector, and denote the length of the vector as .
Step 5. Perform the initialization of the CPSO’s swarm, whose length is .
Step 6. Choose the fitness evaluation of CPSO as the mean squared error (MSE):
Step 7. Execute the optimization process of the CPSO and obtain the best swarm.
Step 8. Reshape the best swarm into four parameters of the network, and then reconstruct the BPANN, consequently named as CPSOBPANN, by using these optimal matrices and vectors.
Step 9. Use to train the CPSOBPANN to obtain BPANN_{Demand}.
Step 10. Obtain the mapping values of the inputs of through
Step 11. Obtain the forecasting values by postprocessing of the output by
Step 12. Replace the electricity demand with DV and repeat Steps 2–11 to obtain the forecasted values of the DVs.
Step 13. Obtain the forecasted price by
Figure 2 schematically shows the entire process of the new model.
5. Numerical Results
In this section, the electricity price and demand will be used to test the models of CPSOBDBPANN, BDBPANN, CPSOBPANN, and BPANN. To begin, the source of data, the forecasting principle, and the cases that must be studied are introduced. Next, cases 1 and 2 are researched specifically with data provided in tables and figures. Finally, the remaining cases are studied, and, consequently, evaluation of the models is performed by comparing the forecasting accuracy, calculation time, and stability.
5.1. Data Selection, Forecasting Principles, and Case Studies
In Data Selection section, the source of the electricity data and certain properties of the data will be briefly introduced. Because a comparison is performed among these models, it is necessary to establish certain principles, which will be mentioned in the section Forecasting Principles. Eventually, to obtain the overall forecasting effectiveness of each model, this paper chooses five cases to illustrate the overall forecasting results calculated by the four models considered.
5.1.1. Data Selection
The electricity price and demand of Victoria of Australia in 2008 are chosen as the database under consideration in this paper. In each day of this database, there are 48 observation points, each representing halfhours. Figure 3 shows the schematic diagram of the electricity price, and Figure 4 shows the schematic diagram of the demand.
It is obvious that demand data are more regular than the electricity price data.
5.1.2. Forecasting Principles
Principle 1. This new forecasting model, whose lifespan is one day, will be updated by reconstructing and training new samples when it has provided 48 forecasting values.
Principle 2. The fourhour points will be used to forecast the price of the next halfhour.
Principle 3. The number of neurons in the hidden layer of the CPSOBPANN is 3.
Principle 4. The points of the previous 21 days of the target day, a total of 1008 halfhour values, construct the training set.
Principle 5. In this paper, the forecasting result of the BPANN is the best result among the 10 times the forecasting results are obtained; this best BPANN is abbreviated as bBPANN.
Principles 1–4 are schematically shown in Figure 5.
5.1.3. Study Cases
There are five cases in this paper. The purpose of case 1 is to explain the proposed model by forecasting only one day specifically. The other cases concentrate on the forecasting effectiveness of the proposed model and a comparison among the new CPSOBDBPANN and CPSOBPANN and the original bBPANN. The evaluation criterion of the models is mean absolute percent error (MAPE), which can be expressed as And it must be pointed out that elements of all columns except the first column are MAPEs which is dimensionless in Tables 4, 5, 6, 7, and 8. The details of each case are presented in Table 1.

5.2. Study of Case 1
In this case, each step, as mentioned in Section 4, will be clearly presented through figures and tables to understand the main concept of this proposed model. First, electricity price, demand, and the value of DV (as described in Section 3) are presented in Figure 6.
The electricity demand clearly has more regularity in the data than those of price and DV. In addition, the fact that the volatility of price and DV do not exhibit distinct differences is also observed. Next, the results of the BDBPANN will be presented in Table 2 and Figure 7.

In Table 2, units of demand and price are Hmk/h and $/Mwh, respectively. The symbol “” represents the time spots, a total of 27 points, for which BDBPANN outperforms bBPANN. The symbol “†” represents the reverse situation to that of the symbol “.” The symbol “‡” represents the results of the average and the variance of each column (except the first column). Specifically, in these 48 time spots, the best forecasting MAPE of demand is 0.05% at 10:30 of this day, and the worst MAPE is 6.06% at 1:00. The average value, the variance, and average forecasting MAPE of the electricity demand obtained through bBPANN are 5645.0323, 555.27757, and 1.38%, respectively. Similarly, the best forecasting value of the DV whose MAPE is 1.25% is at 11:30 of June 9 in 2008, and the worst performance of forecasting occurred at 2:30. The average value, variance, and average forecasting MAPE of the DV which bBPANN calculates are 222.68, 88.982, and 11.24%, respectively. To make a comparison between bBPANN and BDBPANN, the output of bBPANN, that is, the price forecasting values obtained by only using bBPANN, is also presented in Table 2. 0.52% is the best MAPE of bBPANN at 16:30, and 42.39% is the worst MAPE obtained when utilizing bBPANN to forecast electricity price occurred at 2:30. In addition, the average value, variance, and average forecasting MAPE of the price are 30.903, 16.482, and 15.30%, respectively. Furthermore, the best performance of BDBPANN occurred at 20:30, with a forecasting MAPE of 0.56%; meanwhile, 39.00% is the worst MAPE of this new forecasting method. The average forecasting MAPE of the price calculated by BDBPANN is 13.25%, which is less than that of the classical method.
Figure 7 shows the forecasting results of bBPANN and BDBPANN and the actual electricity price. It is clear that BDBPANN is superior to BPANN because the green points (forecasted values of BDBPANN) are closer to the blue curve (actual price) than the red rectangles.
Next, the consequence of using CPSOBDBPANN is described as follows. In Table 3, units of demand and price are Hmk/h and $/Mwh, respectively. The forecasting MAPE of the demand obtained by CPSOBPANN ranges from 0.12% at 15:00 to 4.55% at 1:00, and the CPSOBPANN’s average forecasting MAPE of the demand is 1.23%, which is less than 1.38%, the result of the bBPANN. At 22:30, the best DV forecasted value, whose MAPE is 0.009%, appears. The worst MAPE of DV reaches 41.94% at 8:30, with the average value of the MAPE of DV calculated by CPSOBPANN of 10.79%, which is still less than the 11.24% obtained using bBPANN only. In the end, the final model, CPSOBDBPANN, is used to forecast the electricity prices of June 9 in 2008. Additionally, to enable a comparison between CPSOBPANN and CPSOBDBPANN, the results of CPSOBPANN, that is, to calculate the forecasting values of price by only using CPSOBPANN, are also presented in Table 2. 0.23% is the best MAPE of CPSOBPANN at 8:00, and 42.64% is the worst MAPE of the CPSOBPANN used to forecast electricity price, which occurred at 2:30. In addition, the average MAPE CPSOBPANN is 14.18%, which is better than that of the bBPANN of 15.30%. In the end, the final model in this paper is used to forecast the electricity price of June 9 in 2008. At 10:00 on this day, the new model yields the best forecasting MAPE, which is only 0.36%. On the contrary, at 2:30, the worst performance of CPSOBDBPANN, with a MAPE of 43.07%, arises. The average MAPE obtained by CPSOBDBPANN of this day is 12.31%, which is less than that calculated by CPSOBPANN. However, there are 28 time spots, that is, the symbol “” in Table 3, where the effectiveness of CPSOBDBPANN is better than that of CPSOBPANN. Figure 8 graphically shows the results presented in Table 3.






Overall, it is not difficult to determine that CPSOBDBPANN has the best effectiveness among the models of bBPANN, BDBPANN, CPSOBPANN, and CPSOBDBPANN. Further, the fact that BDBPANN outperforms CPSOBPANN, whose calculating time is much longer than that of BDBPANN, explains why BDBPANN is more efficient than CPSOBPANN, which is an optimal artificial neural network under the same conditions, including calculating capacity and language of debugging.
5.3. Study of Case 2
The primary objective of this part is testing the effectiveness of models in spring, rather than explaining the details of the methods, as was performed in case 1. This section will present the effectiveness of each method, as measured by MAPE. Tables 4 and 5 present a comparison between the bBPANN method and the CPSOBDBPANN and that between CPSOBPANN and BDBPANN. It is obvious that CPSOBDBPANN performs better than bBPANN in all of the spring week except March 11. In March 10, March 13, March 14, and March 16, BDBPANN is more efficient than CPSOBPANN, and the overall effectiveness of BDBPANN is still better than that of CPSOBPANN in this randomly selected spring week. The results indicate that this sample transformation, named bivariate division, of the input of bBPANN is more effective than a complex process using an artificial intelligent algorithm, CPSO in this paper, to obtain the best initial parameters of the network.
Specifically, on March 10, CPSOBPANN exhibited its best MAPE, 0.18%, at 5:30 and obtains its worst one, 103.16%, at 17:00. 1.16% is the minimal MAPE of BDBPANN at 9:00; in contrast, at 14:30, BDBPANN exhibits its maximum MAPE, 56.10%, of this day. The best MAPE of another model, CPSOBDBPANN, is 0.0002% at 8:30 and the worst MAPE for this model reaches 46.52% at 14:30 on this spring day. On the next day of this week, March 11, CPSOBPANN exhibits its best forecasting performance, with a MAPE of 0.02% at 12:30, and the method exhibited a maximum MAPE, 45.55%, at 17:00. Meanwhile, the smallest MAPE of BDBPANN is 0.09% at 22:00, and its maximum one is 42.89% at 7:00. 0.17% is the best forecasting error by using CPSOBDBPANN, which exhibited its worst MAPE of 38.52% at 7:00. On March 12, the MAPE of CPSOBPANN ranges from 0.04% at 9:00 to 35.03% at 7:00; that of BDBPANN reaches its smallest MAPE value, 0.02%, at 11:30 and its maximum one, 43.40%, at 17:00; and CPSOBDBPANN exhibited its best forecasting effectiveness, with a MAPE of 0.2%, at 19:30, and its worst one, 34.21%, at 17:00. On the fourth day of this spring week, bBPANN optimized by the CPSO algorithm obtains its minimal MAPE, whose value is 0.08% at 3:30 and its maximum one, 74.99%, at 19:30; the forecasting MAPE of BDBPANN changes from its smallest value 0.34% at 17:00 to its largest one 62.31% at 15:00 of this day; the final proposed approach, CPSOBDBPANN, exhibited its best forecasting effectiveness at 4:00, with a MAPE of 0.23%, and its worst effectiveness at 15:00, with a MAPE of 56.86%. On March 14, CPSOBPANN exhibited its minimum MAPE of this day of 1.59%, at 21:30, and its largest MAPE of 178.85%, at 19:30; simultaneously, the MAPE of BDBPANN and CPSOBDBPANN ranges from 3.08% at 5:30 and 0.73%, respectively, at 22:00 up to 71.38% and 71.74% for BDBPANN and CPSOBDBPANN, respectively, at 15:00. For the next day, March 15, 0.0003% is the minimum MAPE forecasted by CPSOBPANN at 3:00, and 52.19% at 21:30 is the maximum one; BDBPANN exhibited its best forecasting performance with a MAPE of 0.36% at 6:00 and its worst one with a MAPE of 43.13% at 21:30; using CPSOBDBPANN, its minimum MAPE, 0.34%, and its maximum one, 41.14%, appear at 22:30 and 21:30, respectively. On the last day of this week, the minimum MAPE of CPSOBPANN is 0.18% at 4:00 and the largest one of this method is 62.99% at 14:30; simultaneously, the MAPE of both of the new models, BDBPANN and CPSOBDBPANN, changes from 0.06% at 4:30 and 13:00, respectively, to values of 67.56% and 62.48% for BDBPANN and CPSOBDBPANN, respectively, at 14:30. The results indicate that BDBPANN and CPSOBDBPANN both exhibit better forecasting effectiveness than does CPSOBPANN, especially in the aspect of stability, considering that both BDBPANN and CPSOBDBPANN do not exhibit an extreme MAPE beyond 100%. Figures 9 and 10 show the results of these 4 forecasting methods through the comparison with the actual price and errors of the forecasting values. Thus, the CPSOBDBPANN and BDBPANN methods outperform CPSOBPANN and the original bBPANN.
5.4. Results Analysis and Models Evaluation
In this section, the results of the proposed models of cases 3–5 are presented in Tables 6, 7, and 8. Next, each season’s average MAPE of CPSOBDBPANN, BDBPANN, CPSOBPANN, and bBPANN is presented in Table 9. Simultaneously, the forecasting values of these approaches are also illustrated in pictures. Finally, each of the models will be evaluated through three aspects: forecasting accuracy, calculating time, and stability.

5.4.1. Results Analysis
Table 6 collects the forecasting values calculated by CPSOBDBPANN and BDBPANN. Each column refers to result of each method on different days of summer. For June 9, the lowest MAPE is 0.36% at 10:00 when using BDBPANN, which also exhibited the highest MAPE, 43.70%, at 2:30 on this day. The forecasting MAPE of BDBPANN on June 10 ranges from 0.02% at 3:00 to 44.62% at 8:30. On June 11, the MAPE values of BDBPANN ranged from 0.04% at 15:30 to 41.50% at 1:30. On June 12 of this summer week, CPSOBDBPANN has the best performance, whose MAPE is 0.04% at 16:00, while BDBPANN has the worst one, whose MAPE is 35.42% at 6:00. The MAPE of the fifth day in this week, June 13, changes from 0.07% for CPSOBDBPANN at 15:30 to 32.36% for BDBPANN at 17:00. On June 14, the lowest forecasting MAPE of 0.02% for BDBPANN occurred at 14:00 and CPSOBDBPANN exhibited the worst forecasting effectiveness with a MAPE of 31.51% at 00:00. For the last day of this summer week, the MAPE ranges from 0.001% at 13:00 to 49.06% at 6:30, both of which are for BDBPANN.
Table 7 shows the MAPE computed through using CPSOBDBPANN and BDBPANN to forecast one week of autumn’s electricity price. For September 15, the minimum MAPE is 0.14% at 13:00 when using BDBPANN to forecast price, while CPSOBDBPANN exhibited the maximum MAPE of 35.03% at 11:30. The forecasting MAPE of September 16 changes from 0.12% obtained by CPSOBDBPANN at 10:00 to 45.48% obtained by BDBPANN at 18:00. On September 17, the lowest MAPE value of 0.08% at 14:00 and the highest MAPE of 157.45% at 3:00 were both for BDBPANN. On September 18 of this week, CPSOBDBPANN exhibited the lowest MAPE of 0.19% at 11:30, while its highest MAPE, 29.32% at 6:30, is also the highest in this day. The MAPE of the fifth day, September 19, in this autumn week ranges from 0.96% for CPSOBDBPANN at 5:00 to 27.53% for BDBPANN at 22:00. On September 20, the lowest forecasting MAPE of 0.08% for CPSOBDBPANN was observed at 14:30, while the optimized BDBPANN model exhibited the maximum MAPE of 45.91% at 9:00 of this day. For the last day of this week, the MAPE changes from 0.28% at 9:30 to 43.43% at 23:30, both for BDBPANN.
Table 8 demonstrates the forecasting values calculated by CPSOBDBPANN and BDBPANN. Each column refers to the result of each method on different days of winter. For December 8, the lowest MAPE is 0.06% at 0:00 when using CPSOBDBPANN, while BDBPANN exhibits the highest MAPE of 35.00% at 5:00 on this day. The forecasting MAPE of December 9 ranges from 0.006% at 12:00 to 34.43% at 1:30 which are calculated through CPSOBDBPANN and BDBPANN, respectively. On December 10, the MAPE of CPSOBDBPANN was 0.30% at 0:30 which is the lowest MAPE, while BDBPANN exhibited the highest MAPE of 25.18% at 23:00. On December 11 of this winter week, CPSOBDBPANN has the best performance, whose MAPE is 0.26% at 10:30, while BDBPANN exhibited the worst MAPE of 20.98% at 4:30. The MAPE of the fifth day in this week, December 12, changes from 0.20% at 16:00 to 21.14% at 6:30, with both computed by BDBPANN. On December 13, the lowest forecasting MAPE is 0.07%, calculated by the CPSOBDBPANN at 16:30, and CPSOBDBPANN exhibited the worst forecasting effectiveness, with a MAPE of 23.64% at 10:00 on this day. For the last day of this winter week, the MAPE ranges from 0.17% for CPSOBDBPANN at 18:00 to 125.60% for BDBPANN at 3:00.
Overall, in summer, the average forecasting MAPE of the CPSOBDBPANN method is 9.67%, which is less than that of BDBPANN; the forecasting MPAE of BDBPANN in autumn, 10.72%, is larger than that of CPSOBDBPANN, 10.38%. Similarly, in winter, CPSOBDBPANN forecasts the electricity price more accurately than the original method, BDBPANN; the MAPE of CPSOBDBPANN of 7.40% is less than the MAPE of BDBPANN of 7.74%. Overall, CPSOBDBPANN outperforms BDBPANN 12 times: June 10, June 11, June 12, and June 13 in the summer, September 15, September 16, September 17, and September 20 in autumn, and December 8, December 11, December 13, and December 14 in winter. In contrast, BDBPANN outperforms CPSOBDBPANN 9 times. However, as summarized above, CPSOBDBPANN is a better forecasting model than BDBPANN. Figures 11, 12, and 13 show the results of the new approaches, the actual values, and the forecasting errors.
5.4.2. Models Evaluation
In the above portion of 5.4, the results of both proposed forecasting models of cases 3, 4, and 5 were presented in tables and figures. These results indicate that CPSOBDBPANN outperforms BDBPANN in the aspect of accuracy. Furthermore, to test the forecasting effectiveness of CPSOBDBPANN and BDBPANN, this paper compares both of them with bBPANN and CPSOBPANN in the aspects of forecasting accuracy, calculation time, and stability. In this study, the forecasting accuracy is measured by MAPE, and the index of stability of the models is the number of points of the forecasting MAPE that exceed 100%; that is, a model can be regarded as a relatively stable model when the index of stability is relatively low. The unit of calculation time is one second. Table 9 presents the results of the evaluation of the models of bBPANN, CPSOBPANN, BDBPANN, and CPSOBDBPANN.
From Table 9, it is obvious that CPSOBDBPANN is the best forecasting model when considering accuracy. In the aspect of calculation time, bBPANN and BDBPANN significantly outperform the other models. Finally, the last row of this table indicates that the proposed models are better than bBPANN and CPSOBPANN regarding forecasting stability. As a result, to forecast the electricity price efficiently, either BDBPANN or CPSOBDBPANN should be selected as the forecasting model. Specifically, when accuracy is the only concern, CPSOBDBPANN should be used to forecast the electricity price. However, when the cost of calculation is the only concern, BDBPANN is the best model because its accuracy is higher than that of BPANN, the forecasting accuracy of which must be less than that of bBPANN (by the definition of bBPANN).
6. Our Contribution
(i)To combine historical price and demand to improve the forecasting of the electricity price, this paper proposed a new variable, known as DV, through a division of one variable by the other, rather than simply regarding them as inputs of a network. In this approach, multiplication of the forecasted DV and the demand is taken as the final forecasted price, which consequently improves the accuracy of forecasting price than that using the original BPANN.(ii)The CPSO algorithm, BPANN, and the concept of bivariate division are combined to form the new models: BDBPANN and CPSOBDBPANN.(iii)This paper evaluated the models through three aspects: forecasting accuracy, calculating time, and stability. The evaluation indicated that CPSOBDBPANN is a relatively better model when only considering accuracy and that BDBPANN is a comparatively better model when the cost of calculation should be low, while the accuracy remains high.
7. Conclusions
In this paper, the BDBPANN method is creatively proposed by combining the concept of bivariate division with BPANN to forecast electricity price more accurately and stably than BPANN does. In four seasons of 2008, the average MAPE of BDBPANN is 10.36%, which is less than those of BPANN and CPSOBPANN. Note that BDBPANN performs better than BPANN and CPSOBPANN in each season. Next, the CPSO algorithm was applied to optimize the initial parameters of BDBPANN to improve the forecasting accuracy. This optimized forecasting model, named CPSOBDBPANN, forecasted the price with an average MAPE, 10.02%, in the weeks of the four seasons considered. Additionally, the CPSOBDBPANN method outperforms the other approaches in forecasting accuracy. Finally, the evaluation of the models of BPANN, BDBPANN, CPSOBPANN, and CPSOBDBPANN was performed. This evaluation involved tests regarding accuracy, calculation time, and stability. For the aspect of accuracy, CPSOBDBPANN is the best model. The BPANN model, which exhibits an average calculation time of 3.34 seconds, has huge advantages compared with the other models regarding the calculation time. The last test, stability, indicated that the proposed methods, both of which have an index of stability of 0.25, are more stable than the original BPANN and the CPSOBPANN models, with indices of stability of 3 and 1.5, respectively. These results illustrate that CPSOBDBPANN is a superior model to the others when only concerning accuracy. And BDBPANN outperforms CPSOBDBPANN, CPSOBPANN, and BPANN when calculation time is the crucial factor.
Highlights
The highlights of this paper are as follows: skillful transformation of the electricity price and demand into a new variable, known as DV, rather than only regarding both of them as inputs of the network; utilization of chaos particle swarm optimization to obtain the relatively improved weights and thresholds of the network, rather than the random ones; proposal of two different forecasting strategies for the user of the model.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgment
The work was supported by the National Natural Science Foundation of China (Grant no. 71171102).
References
 S. Anbazhagan and N. Kumarappan, “Dayahead deregulated electricity market price classification using neural network input featured by DCT,” International Journal of Electrical Power and Energy Systems, vol. 37, no. 1, pp. 103–109, 2012. View at: Publisher Site  Google Scholar
 S. K. Aggarwal, L. M. Saini, and A. Kumar, “Electricity price forecasting in deregulated markets: a review and evaluation,” International Journal of Electrical Power and Energy Systems, vol. 31, no. 1, pp. 13–22, 2009. View at: Publisher Site  Google Scholar
 H.T. Pao, “Forecasting electricity market pricing using artificial neural networks,” Energy Conversion and Management, vol. 48, no. 3, pp. 907–912, 2007. View at: Publisher Site  Google Scholar
 X. R. Li, C. W. Yu, S. Y. Ren, C. H. Chiu, and K. Meng, “Dayahead electricity price forecasting based on panel cointegration and particle filter,” Electric Power Systems Research, vol. 95, pp. 66–76, 2013. View at: Publisher Site  Google Scholar
 J. Zhang, Z. Tan, and S. Yang, “Dayahead electricity price forecasting by a new hybrid method,” Computers and Industrial Engineering, vol. 63, no. 3, pp. 695–701, 2012. View at: Publisher Site  Google Scholar
 Z. Tan, J. Zhang, J. Wang, and J. Xu, “Dayahead electricity price forecasting using wavelet transform combined with ARIMA and GARCH models,” Applied Energy, vol. 87, no. 11, pp. 3606–3610, 2010. View at: Publisher Site  Google Scholar
 N. Amjady and F. Keynia, “Dayahead price forecasting of electricity markets by a new feature selection algorithm and cascaded neural network technique,” Energy Conversion and Management, vol. 50, no. 12, pp. 2976–2982, 2009. View at: Publisher Site  Google Scholar
 V. Vahidinasab, S. Jadid, and A. Kazemi, “Dayahead price forecasting in restructured power systems using artificial neural networks,” Electric Power Systems Research, vol. 78, no. 8, pp. 1332–1342, 2008. View at: Publisher Site  Google Scholar
 A. I. Arciniegas and I. E. Arciniegas Rueda, “Forecasting shortterm power prices in the Ontario Electricity Market (OEM) with a fuzzy logic based inference system,” Utilities Policy, vol. 16, no. 1, pp. 39–48, 2008. View at: Publisher Site  Google Scholar
 S. Chakravarty and P. K. Dash, “Dynamic filter weights neural network model integrated with differential evolution for dayahead price forecasting in energy market,” Expert Systems with Applications, vol. 38, no. 9, pp. 10974–10982, 2011. View at: Publisher Site  Google Scholar
 S. Bordignon, D. W. Bunn, F. Lisi, and F. Nan, “Combining dayahead forecasts for British electricity prices,” Energy Economics, vol. 35, pp. 88–103, 2013. View at: Publisher Site  Google Scholar
 H. Shayeghi and A. Ghasemi, “Dayahead electricity prices forecasting by a modified CGSA technique and hybrid WT in LSSVM based scheme,” Energy Conversion and Management, vol. 74, pp. 482–491, 2013. View at: Publisher Site  Google Scholar
 S. Anbazhagan and N. Kumarappan, “Dayahead deregulated electricity market price forecasting using neural network input featured by DCT,” Energy Conversion and Management, vol. 78, pp. 711–719, 2014. View at: Google Scholar
 D. Niu, D. Liu, and D. D. Wu, “A soft computing system for dayahead electricity price forecasting,” Applied Soft Computing Journal, vol. 10, no. 3, pp. 868–875, 2010. View at: Publisher Site  Google Scholar
 W.M. Lin, H.J. Gow, and M.T. Tsai, “An enhanced radial basis function network for shortterm electricity price forecasting,” Applied Energy, vol. 87, no. 10, pp. 3226–3234, 2010. View at: Publisher Site  Google Scholar
 X. Yan and N. A. Chowdhury, “Midterm electricity market clearing price forecasting: a hybrid LSSVM and ARMAX approach,” International Journal of Electrical Power and Energy Systems, vol. 53, no. 1, pp. 20–26, 2013. View at: Publisher Site  Google Scholar
 X. Yan and N. A. Chowdhury, “Midterm electricity market clearing price forecasting: a multiple SVM approach,” International Journal of Electrical Power and Energy Systems, vol. 58, pp. 206–214, 2014. View at: Google Scholar
 W.M. Lin, H.J. Gow, and M.T. Tsai, “Electricity price forecasting using enhanced probability neural network,” Energy Conversion and Management, vol. 51, no. 12, pp. 2707–2714, 2010. View at: Publisher Site  Google Scholar
 D. Singhal and K. S. Swarup, “Electricity price forecasting using artificial neural networks,” International Journal of Electrical Power and Energy Systems, vol. 33, no. 3, pp. 550–555, 2011. View at: Publisher Site  Google Scholar
 B. D. Rodrigues and M. J. Stevenson, “Takeover prediction using forecast combinations,” International Journal of Forecasting, vol. 29, no. 4, pp. 628–641, 2013. View at: Publisher Site  Google Scholar
 N. An, W. Zhao, J. Wang, D. Shang, and E. Zhao, “Using multioutput feedforward neural network with empirical mode decomposition based signal filtering for electricity demand forecasting,” Energy, vol. 49, no. 1, pp. 279–288, 2013. View at: Publisher Site  Google Scholar
 X. W. Li, S. J. Cho, and S. T. Kim, “Combined use of BP neural network and computational integral imaging reconstruction for optical multipleimage security,” Optics Communications, vol. 315, pp. 147–158, 2014. View at: Google Scholar
 R. HechtNelson, “Kolmogorov's mapping neural network existence theorem,” in Proceedings of 1st IEEE Annual International Conference on Neural Networks, pp. 11–13, IEEE Press, San Diego, Calif, USA, 1987. View at: Google Scholar
 Z.H. Guo, J. Wu, H.Y. Lu, and J.Z. Wang, “A case study on a hybrid wind speed forecasting method using BP neural network,” KnowledgeBased Systems, vol. 24, no. 7, pp. 1048–1056, 2011. View at: Publisher Site  Google Scholar
 J. Wang, Z. Sheng, B. Zhou, and S. Zhou, “Lightning potential forecast over Nanjing with denoised soundingderived indices based on SSA and CSBP neural network,” Atmospheric Research, vol. 137, pp. 245–256, 2014. View at: Google Scholar
Copyright
Copyright © 2014 Zhilong Wang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.