CN108197748B - A People Flow Prediction Method Based on Thinking Evolution Algorithm - Google Patents

Abstract

The invention discloses a people flow prediction method based on a thought evolution algorithm, which is based on a neural network prediction model, comprehensively considers the factors influencing the people flow of a power supply business hall, takes the input data of the prediction model as the corresponding time point, the weather of the day and whether the day is a working day, and outputs the people flow prediction value at the expected time. The method optimizes the traditional neural network structure through a thinking evolution algorithm, optimizes the hidden layer weight and the threshold through convergence and differentiation operations, and reduces the prediction error caused by random generation of the hidden layer weight and the threshold, thereby constructing a prediction model with higher precision, improving the people flow prediction level of the power supply business hall and providing more accurate scientific basis for the scheme formulation of the power supply business hall for improving the self service level.

Description

A People Flow Prediction Method Based on Thinking Evolution Algorithm

technical field

The invention relates to a human flow prediction method, which belongs to the technical field of intelligent prediction.

Background technique

With the gradual integration of power enterprise business into a deeper market economic environment, improving power marketing business with high-quality services has become a necessary work for current power enterprises to carry out daily operation and management. As a window unit directly facing power users, the service level of the power supply business hall directly affects the impression of the power customers on the enterprise.

The flow of people in the power supply business hall is the core reference for the staffing and equipment of the power supply business hall, and it is a key factor affecting the service quality of the power supply business hall itself. Although traditional neural network can be used for numerical prediction of human flow, its prediction accuracy is not high and still needs to be improved. Therefore, a more accurate modeling method for human flow prediction needs to be designed.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to propose a new method for predicting the flow of people, aiming at the defects involved in the background technology, so as to improve the prediction accuracy.

The present invention adopts the following technical solutions in order to solve the above technical problems:

A method for predicting human flow based on a thinking evolution algorithm, comprising the following steps:

Step 1: Set the data format of the corresponding time point, the weather data format of the day, and whether it is a working day data format;

Step 2: Construct the model input data feature vector:

E =[ T _i , Wi _, K _i ]

Wherein, the weather of the day at time i is recorded as Wi , whether it is a work diary is K _{i ,} and the corresponding time point is recorded as T _i ;

Step 3: Set parameters for the thinking evolution algorithm, the set parameters include population size, winning sub-population, temporary sub-population, hidden layer, and number of iterations;

Step 4: Perform a convergence operation on the evolutionary thinking algorithm. When the score of the temporary sub-population is lower than the score of the winning sub-population, the convergence operation is stopped, otherwise, the convergence operation is continued.

Step 5: When the convergence operation stops, obtain the optimal weights and thresholds of the BP neural network;

Step 6: Input historical data, train the BP neural network, and obtain a prediction model;

Step 7: Input the corresponding time point to be predicted, the weather of the day, and whether it is a working day into the prediction model, and then the flow of people at the time point to be predicted can be obtained.

The present invention adopts the above technical scheme, and has the following technical effects:

This method is based on the neural network prediction model, optimizes the traditional neural network structure through the evolutionary algorithm of thinking, and obtains a prediction model with higher prediction accuracy, which improves the flow prediction level of the power supply business hall, and improves the service level of the power supply business hall. The formulation of the program provides a more accurate scientific basis.

Description of drawings

FIG. 1 is a flow chart of the present invention.

FIG. 2 is a schematic diagram of the present invention.

Detailed ways

Below in conjunction with accompanying drawing, the technical scheme of the present invention is described in further detail:

It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art and, unless defined as herein, are not to be taken in an idealized or overly formal sense. explain.

The invention provides a human flow prediction modeling method based on a thinking evolution algorithm. The model is based on a BP neural network, and the thinking evolution algorithm is introduced to optimize the structure of the BP neural network, thereby improving the prediction accuracy. The thinking evolution algorithm is an optimization algorithm that evolves continuously through iteration. This method can realize the optimization of the neural network prediction model: optimize the hidden layer weights and thresholds through convergence and alienation operations, and reduce the hidden layer weights and thresholds. The resulting prediction error can be used to build a more accurate prediction model.

As shown in Figure 2, a kind of people flow prediction modeling method based on thinking evolution algorithm of the present invention comprehensively considers the factors affecting the flow of people in the power supply business hall, and takes the input data of the prediction model as the corresponding time point, the weather of the day, Whether it is a working day, the output is the predicted value of people flow at the desired time.

Corresponding time point data format: According to the research and analysis of the business handling situation, the sampling period is 5 minutes. Among them, the processing method for the corresponding time point is: adopt the 24-hour system, and the M hour and N minutes are recorded as ( M + N /60), taking 14:15 minutes as an example, that is, 14+15/60=14.25.

Today's weather data format: According to the degree to which the weather affects travel, it is divided into 5 levels, and each level is assigned a code: (sunny = 1), (cloudy = 2), (light rain = 3), (moderate rain/ Light Snow/Sleet = 4), (Heavy Rain/Snow = 5), (Extremely Heavy Rain/Snow or Typhoon = 6).

Whether it is working day data format: 1 for working diary, 0 for non-working diary: (Mon-Fri = 1), (Sat/Sun = 0).

Construct the input data feature vector of the model: the weather condition at time i is recorded as Wi , the working day is recorded as K _{i ,} and the time point at this time is recorded as T _i . Then the model input feature vector is:

E = [ T _i , Wi _, K _i ].

Set parameters for the mind evolution algorithm. The thinking evolution algorithm is an algorithm that continuously evolves through iteration. The collection of all individuals in each generation during evolution is called a group, and a group is divided into several subgroups. The set parameters include population size, winning subpopulation, temporary subpopulation, hidden layer, and number of iterations.

Perform a convergence operation on the evolutionary algorithm of thinking (within the sub-group, the process of individuals competing to become the winner is convergence), when the score of the temporary sub-population is lower than the score of the winning sub-population, the convergence operation stops, otherwise, it continues. Convergence operation.

When the convergence operation is stopped, the optimal weights and thresholds of the BP neural network are obtained, that is, the prediction model reaches the optimal structure.

Input historical data, train BP neural network, and get a prediction model.

Input the time point to be predicted, the weather of the day, and whether it is a working day into the prediction model, and the flow of people at the expected time of the power supply business hall can be obtained.

The model is based on the BP neural network, as shown in Figure 1. The structure of the BP neural network is optimized by introducing the evolutionary algorithm of thinking, thereby improving the prediction accuracy.

The above are only some embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the principles of the present invention, several improvements and modifications can be made. It should be regarded as the protection scope of the present invention.

Claims (5)

1. a kind of people flow prediction method based on thinking evolution algorithm, is characterized in that, comprises the following steps: Step 1: Set the data format of the corresponding time point, the weather data format of the day, and whether it is a working day data format; Step 2: Construct prediction model input data feature vector: E =[ T _i , Wi _, K _i ] Wherein Wi is the weather of the day at time _i , whether it is a work diary is Ki _, and the corresponding time point is recorded as Ti _; Step 3: Set parameters for the thinking evolution algorithm. The set parameters include population size, winning sub-population, temporary sub-population, hidden layer, and number of iterations; Step 4: Carry out the convergence operation on the evolutionary thinking algorithm. When the score of the temporary sub-population is lower than the score of the winning sub-population, the convergence operation is stopped, otherwise, the convergence operation is continued; Step 5: When the convergence operation stops, obtain the optimal weights and thresholds of the BP neural network; Step 6: Input historical data, train the BP neural network, and obtain a prediction model; Step 7: Enter the corresponding time point to be predicted, the weather of the day, and whether it is a working day into the prediction model, and then the flow of people at the time to be predicted can be obtained. 2 . The method for predicting the flow of people based on an evolutionary thinking algorithm according to claim 1 , wherein the sampling period of the data is 5 minutes. 3 . 3. a kind of people flow forecasting method based on evolutionary algorithm of thinking according to claim 1, is characterized in that, in step 1, for the data format of corresponding time point: adopt 24-hour system, and N minutes are recorded as ( M when M. + N /60). 4. a kind of people flow prediction method based on evolutionary algorithm of thinking according to claim 1, is characterized in that, in step 1, for the weather data format of the day is: according to the degree that weather affects travel, it is divided into K levels, Assign a code to each level. 5. a kind of people flow prediction method based on evolutionary algorithm of thinking according to claim 1, is characterized in that, in step 1, whether it is working day data format is: working diary is 1, non-working diary is 0.

Images