JP6927593B2 - Information processing equipment and data structure - Google Patents

Description

The present invention relates to an information processing device, a distribution achievement / failure prediction method, and a distribution achievement / failure prediction program, and more particularly to an information processing device for creating a delivery schedule, a distribution achievement / failure prediction method, and a distribution achievement / failure prediction program.

In recent years, the number of redelivery cases has increased in the home delivery business. Here, the redelivery rate accounts for about 20% of the total delivery (announced by the Ministry of Land, Infrastructure, Transport and Tourism in December 2012), and labor costs and redelivery costs are large losses in the home delivery business.

Therefore, Patent Document 1 discloses a technique relating to a delivery planning system for improving delivery efficiency. The delivery planning system according to Patent Document 1 first manages the time when delivery was performed in the past and the home / absence of the delivery destination in the delivery record database, and has a category showing the customer's behavior pattern and home time pattern. To classify each customer into multiple categories by comparing. Then, the delivery planning system calculates the delivery time within the estimated home time from the order status and the category to which the customer belongs, and creates a delivery schedule.

Further, Patent Document 2 discloses a technique relating to a system for calculating the probability of staying at home using information on the amount of electric power used. The system according to Patent Document 2 calculates the probability of staying at home using the maximum and minimum values of the user's power consumption.

Japanese Unexamined Patent Publication No. 2013-170050 Japanese Unexamined Patent Publication No. 2012-181789

However, it cannot be said that the techniques according to Patent Documents 1 and 2 have sufficient accuracy in predicting the success or failure of delivery. For example, in Patent Document 1, it can be said that the accuracy of the probability of staying at home is insufficient for a time zone in which there is no delivery record. Further, in Patent Document 2, if the amount of electric power used is large, it is uniformly determined that the person is at home, but depending on the electric device, the electric power may be used even when he / she is absent, and it can be said that the accuracy of the probability of being at home is still insufficient.

The present invention has been made to solve such a problem, and an object of the present invention is to provide an information processing device for predicting the success or failure of delivery with high accuracy, a delivery achievement / failure prediction method, and a delivery achievement / failure prediction program. And.

The information processing device according to the first aspect of the present invention is
Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is A pattern information extraction unit that extracts pattern information indicating a change tendency for each success or failure,
A prediction unit that compares the power consumption information at an arbitrary delivery destination with the extracted pattern information and predicts the success or failure of delivery for each time zone at the arbitrary delivery destination.
To be equipped.

The method for predicting the achievement or failure of distribution according to the second aspect of the present invention is
Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is Pattern information showing a change tendency is extracted for each success or failure, and
By comparing the power consumption information at an arbitrary delivery destination with the extracted pattern information, the success or failure of delivery for each time zone at the arbitrary delivery destination is predicted.
To be equipped.

The allocation achievement prediction program according to the third aspect of the present invention is
Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is The process of extracting pattern information indicating a change tendency for each success or failure, and
A process of comparing the power consumption information at an arbitrary delivery destination with the extracted pattern information to predict the success or failure of delivery for each time zone at the arbitrary delivery destination.
Let the computer run.

INDUSTRIAL APPLICABILITY According to the present invention, it is possible to provide an information processing device for predicting the success or failure of delivery with high accuracy, a delivery achievement / failure prediction method, and a delivery achievement / failure prediction program.

It is a block diagram which shows the structure of the information processing apparatus which concerns on Embodiment 1 of this invention. It is a flowchart which shows the flow of the distribution achievement / failure prediction processing which concerns on Embodiment 1 of this invention. It is a block diagram which shows the structure of the delivery schedule management server which concerns on Embodiment 2 of this invention. It is a flowchart which shows the flow of the delivery schedule creation processing which concerns on Embodiment 2 of this invention. It is a flowchart which shows the flow of the pattern information extraction processing which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the relationship between the power consumption and the delivery record which concerns on Embodiment 2 of this invention. It is a figure for demonstrating the concept of the extraction of the power consumption pattern which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the delivery success electric power pattern which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the delivery failure electric power pattern which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the normalized delivery success electric power pattern which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the normalized delivery failure power pattern which concerns on Embodiment 2 of this invention. It is a figure for demonstrating the concept of matching of the power consumption pattern which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the power consumption of the home rate calculation target which concerns on Embodiment 2 of this invention. It is a figure which shows the example of the home rate information of a specific consumer which concerns on Embodiment 2 of this invention. It is a flowchart which shows the flow of the latest home rate calculation processing which concerns on Embodiment 3 of this invention. It is a figure which shows the example of the most recent power consumption which concerns on Embodiment 3 of this invention.

Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. In each drawing, the same elements are designated by the same reference numerals, and duplicate description will be omitted as necessary for the sake of clarity of description.

<Embodiment 1>
FIG. 1 is a block diagram showing a configuration of an information processing apparatus 1 according to a first embodiment of the present invention. The information processing device 1 includes a pattern information extraction unit 11 and a prediction unit 12. The pattern information extraction unit 11 extracts pattern information (described later) based on the delivery record information 21 and the power consumption information 22. Here, the delivery record information 21 is information including success / failure and success / failure time of delivery to a predetermined delivery destination. Here, the "success or failure of delivery" is information indicating that the delivery to the predetermined delivery destination was successful or unsuccessful (that is, the person was absent and the delivery was not completed). Further, the "delivery success / failure time" is information associated with the "delivery success / failure", and indicates the time when the delivery to the predetermined delivery destination is successful or the time when the delivery is unsuccessful. The delivery record information 21 is, for example, information accumulated by the transportation company. Further, the power consumption information 22 is information indicating the time-series power consumption at a predetermined delivery destination. The power consumption information 22 is, for example, information accumulated by the electric power company. The delivery record information 21 and the power consumption information 22 are acquired from a database (DB) or the like owned by the transportation company and the electric power company, respectively, via a network (not shown), or the acquired information is an information processing device. It may be stored in the storage device (not shown) in 1.

In addition, the pattern information extraction unit 11 extracts pattern information for each success or failure. Here, the pattern information is information indicating a change tendency of power consumption in a time zone around the success / failure time. Then, the pattern information is extracted for each of the case where the delivery is successful and the case where the delivery is unsuccessful. It should be noted that the pattern information includes information similar in success or failure.

The prediction unit 12 compares the power consumption information 22 at an arbitrary delivery destination with the extracted pattern information, and predicts the success or failure of delivery for each time zone at the arbitrary delivery destination. Here, it does not matter whether or not any delivery destination is the same as the above-mentioned predetermined delivery destination.

FIG. 2 is a flowchart showing the flow of the allocation achievement / failure prediction process according to the first embodiment of the present invention. First, the pattern information extraction unit 11 extracts pattern information indicating a change tendency of power consumption in a time zone around the success / failure time based on the delivery record information 21 and the power consumption information 22 at a predetermined delivery destination (S11). ).

Next, the prediction unit 12 compares the power consumption information 22 at the arbitrary delivery destination with the extracted pattern information, and predicts the success or failure of the delivery for each time zone at the arbitrary delivery destination (S12).

As described above, in the first embodiment, when the change tendency of the electric power consumption is analyzed from the time zone in which the electric power consumption at the predetermined delivery destination has the delivery record, and the delivery is successful. And the pattern information in case of failure is extracted. Then, from the actual power consumption at any delivery destination, for example, a time zone in which a change tendency similar to the pattern information appears is specified. As a result, it is possible to predict with high accuracy the success or failure of future delivery in the time zone specified according to the success or failure of delivery of the pattern information. For example, when a predetermined delivery destination is different from an arbitrary delivery destination, even if the house has no delivery record, if the power consumption information in the house can be obtained, the pattern information in the predetermined delivery destination can be used arbitrarily. It is possible to predict in advance the success or failure of delivery at the delivery destination. In addition, even if a predetermined delivery destination and an arbitrary delivery destination are the same, the success or failure of delivery in a time zone without a delivery record at the same delivery destination is predicted by using pattern information in a time zone with a delivery record. be able to.

<Embodiment 2>
The second embodiment is an embodiment of the first embodiment described above.
FIG. 3 is a block diagram showing the configuration of the delivery schedule management server 100 according to the second embodiment of the present invention. The delivery schedule management server 100 is an example of the information processing device 1 described above, creates and manages a delivery schedule for an arbitrary delivery destination, and transmits the delivery schedule to the mobile terminal 200 owned by the delivery person.

The delivery schedule management server 100 includes a delivery record DB 111, a power consumption DB 112, a power consumption pattern DB 113, a home rate DB 114, a delivery schedule DB 115, a collection unit 121, a pattern information extraction unit 122, and a home rate calculation unit. It includes 123, a delivery schedule creation unit 124, and an output unit 125.

The delivery record DB 111 is a database in which delivery record data indicating the success or failure record of delivery by the shipping company is collected and stored. The delivery record data is at least information in which information indicating the delivery destination (consumer), a flag indicating the success or failure of the delivery, and the time when the delivery succeeded or failed, that is, the success or failure time are associated with each other. The delivery record data is an example of the above-mentioned delivery record information 21.

The power consumption DB 112 is a database collected and stored from a DB of an electric power company or the like for power consumption data indicating the power consumption for each predetermined time measured by a smart meter or the like installed in each household. The power consumption data is associated with at least information indicating a consumer (here, a person corresponding to the delivery destination stored in the delivery record DB 111), the time when the power is measured, and the measured power amount. Information. The power consumption data is an example of the power consumption information 22 described above.

The power consumption pattern DB 113 is a database in which the power consumption patterns analyzed from the delivery record DB 111 and the power consumption DB 112 are stored. Here, it is assumed that the power consumption pattern can be distinguished from the one at the time of successful delivery (delivery success power pattern) and the one at the time of delivery failure (delivery failure power pattern). Further, it is assumed that the power consumption pattern includes an individual power consumption pattern (individual power consumption pattern) and a grouped (normalized) power consumption pattern (normalized power consumption pattern).

The individual power consumption pattern includes, for example, an individual pattern ID, a flag indicating the success or failure of delivery, delivery time information, power pattern information, and a delivery destination. Here, the delivery time information includes, for example, a combination of the delivery date and time, the time before or after the pattern occurs from the delivery date and time (occurrence time), and the end time of the pattern, or instead of the end time of the pattern. Includes the time width of the pattern. Further, the power pattern information is data corresponding to a part of the time zone in the power consumption data. Alternatively, the power pattern information is a combination of the time of difference from the immediately preceding extreme value (maximum value or minimum value), the amount of power (power change amount) of the difference from the immediately preceding extreme value, and the characteristics of the section (described later). be. Further, the normalized power consumption pattern includes, for example, a power pattern ID, a flag indicating success or failure of delivery, power pattern information, generation timing, number of occurrences, delivery success probability, and the like. The power consumption pattern stored in the power consumption pattern DB 113 is an example of the above-mentioned pattern information.

The home-based rate DB 114 is a database in which the home-based rate, which is the probability of being at home for each month, day of the week, and time zone for each consumer analyzed from the power consumption DB 112 and the power consumption pattern DB 113, is stored. The home-based rate is an example of the result of prediction of success or failure of delivery by the above-mentioned prediction unit 12.

The delivery schedule DB 115 is a database in which a delivery schedule including a delivery order, a scheduled delivery time, a delivery route, and the like for an arbitrary delivery destination created based on the home rate is stored.

The delivery destination list DB116 scheduled to be delivered today is a database in which information on the delivery destination scheduled to be delivered today (identification information of the delivery destination, address, telephone number, desired delivery time zone, etc.) is stored among a plurality of delivery destinations. be.

The collecting unit 121 collects the delivery record data from the transportation company or the like and stores it in the delivery record DB 111, and collects the power consumption data from the electric power company or the like and stores it in the power consumption DB 112. The pattern information extraction unit 122 is an example of the pattern information extraction unit 11 described above, and refers to the delivery record DB 111 and the power consumption DB 112, and changes in power consumption in the surrounding time zone for each success / failure time at a predetermined delivery destination. The tendency is analyzed, and the power consumption data of a certain characteristic time zone is extracted as an individual power consumption pattern. At this time, the pattern information extraction unit 122 divides the individual power consumption pattern into a delivery success power pattern at the time of successful delivery and a delivery failure power pattern at the time of delivery failure, and stores the individual power consumption pattern in the power consumption pattern DB 113. Then, the pattern information extraction unit 122 normalizes a plurality of delivery success power patterns and delivery failure power patterns that are similar, and stores them in the power consumption pattern DB 113 as the normalized power consumption patterns. At this time, the pattern information extraction unit 122 divides the normalized delivery success power pattern and the normalized delivery failure power pattern and stores them in the power consumption pattern DB 113. At the same time, the pattern information extraction unit 122 calculates the delivery success degree (delivery success probability) according to the success or failure of the delivery for each normalized power consumption pattern, and stores it in the power consumption pattern DB 113 in association with the normalized power consumption pattern. do.

The relationship between this power consumption pattern and delivery success or failure is that the delivery success rate is high after using a certain power device (when using a microwave oven, eat afterwards, etc.), or a certain power device is used. In this case, the delivery success rate is low immediately before that (such as turning on the lighting after returning home), and there is no dependency between the power equipment and the delivery success rate (standby power such as a refrigerator). The power consumption pattern extraction logic will be described later.

The home rate calculation unit 123 is an example of the above-mentioned prediction unit 12, and refers to the power consumption DB 112 and the power consumption pattern DB 113, and within a typical period (for example, month, day of the week, and time) at a certain delivery destination. Calculate the home-based rate, which is the probability that a consumer (a person who can receive deliveries) is at home at the delivery destination.

The delivery schedule creation unit 124 creates a delivery schedule calculated on the condition of the home rate DB 114 from the delivery destination list DB 116 scheduled to be delivered today, and stores it in the delivery schedule DB 115. The output unit 125 transmits the delivery schedule stored in the delivery schedule DB 115 to the mobile terminal 200 or the like.

The mobile terminal 200 is a mobile terminal possessed by a driver or the like in charge of delivery, and displays a delivery order, a scheduled delivery time, a delivery route, and the like on a screen based on a delivery schedule received from the delivery schedule management server 100.

FIG. 4 is a flowchart showing the flow of the delivery schedule creation process according to the second embodiment of the present invention. First, the collection unit 121 collects delivery record data of a plurality of delivery destinations from a transportation company or the like and stores them in the delivery record DB 111 (S110). In addition, the collection unit 121 collects power consumption data of a plurality of delivery destinations from an electric power company or the like and stores the power consumption data in the power consumption DB 112 (S120). At this time, at least a part of the delivery destinations in the data stored in the delivery record DB 111 and the power consumption DB 112 is assumed to correspond.

Next, the pattern information extraction unit 122 extracts the power consumption pattern (S130). Specifically, FIG. 5 shows a flowchart showing the flow of the pattern information extraction process according to the second embodiment of the present invention. The pattern information extraction process shall be performed for a plurality of delivery destinations common to the delivery record DB 111 and the power consumption DB 112.

First, the pattern information extraction unit 122 acquires the success / failure and success / failure time of delivery to a predetermined delivery destination from the delivery record DB 111 (S131). Next, the pattern information extraction unit 122 acquires the power consumption data of the time zone around the acquired success / failure time from the power consumption data at the predetermined delivery destination in the power consumption DB 112 (S132). Here, FIG. 6 is a diagram showing an example of the relationship between the power consumption and the delivery record according to the second embodiment of the present invention. The area of the alternate long and short dash line in FIG. 6 shows the power consumption data of the time zone around the success / failure time of delivery. That is, matching is performed between the measurement time in the power consumption DB 112 and the success / failure time in the delivery record DB 111.

The pattern information extraction unit 122 calculates the amount of power change within a predetermined range for the time zones before and after the success / failure time, and determines the characteristics of the power change (S133). For example, the pattern information extraction unit 122 detects an extreme value (maximum value or minimum value) of power consumption in a predetermined range along a time series, and calculates the time interval between adjacent extreme values and the amount of power change. do. Then, the pattern information extraction unit 122 determines the characteristics of each time interval based on the time interval and the amount of power change. FIG. 7 is a diagram for explaining the concept of extracting the power consumption pattern according to the second embodiment of the present invention. In FIG. 7, the time interval between some extreme values is omitted.

Here, an example will be given of how to determine the characteristics. For example, when the time interval is less than a predetermined time and the amount of change in power is equal to or more than a predetermined value, the pattern information extraction unit 122 determines that the features are "rapid increase" and "rapid decrease". Further, when the time interval is equal to or longer than a predetermined time and the amount of change in power is less than a predetermined value, the pattern information extraction unit 122 determines that the feature is "no change for a certain period of time". Further, when the time interval is less than a predetermined time, the amount of power change is a predetermined value, and the features "rapid increase" and "rapid decrease" are repeated a certain number of times, the pattern information extraction unit 122 determines that the feature is "periodic change". In other words, it is characteristic that the section where the amount of change in electric power changes significantly, the period when the electric power does not change for a certain period of time, the period and swing width when repeating periodically are detected, and when they exceed a certain threshold value, etc. Extract as a power pattern.

The method of determining the characteristics is not limited to these. For example, a characteristic part such as a sharp curve, a periodic fluctuation, and a change in power consumption at that time may be detected in a change in power consumption in a time series.

Then, the pattern information extraction unit 122 saves each success or failure as an individual power consumption pattern (delivery success power pattern or delivery failure power pattern) based on the order of features and the amount of power change within a predetermined period (S134). Here, FIG. 8 is a diagram showing an example of a successful delivery power pattern according to the second embodiment of the present invention. Further, FIG. 9 is a diagram showing an example of a delivery failure power pattern according to the second embodiment of the present invention. The "occurrence time" in FIGS. 8 and 9 is "-" in the case of the time before delivery (that is, in advance) and "+" in the case of the time after delivery (that is, after) based on the success / failure time. The time of the difference is recorded. As a result, the relationship between the power consumption pattern and the success or failure of delivery is that the delivery success rate is high after using a certain power device (for example, after using a microwave oven, then eating), and the certain power device. When is used, it is possible to catch the tendency that the delivery success rate is low immediately before that (for example, the lighting is turned on after returning home).

When the success / failure of the acquired delivery indicates the success of the delivery, the pattern information extraction unit 122 sets the time zone (advance) before the success / failure time corresponding to the delivery as the surrounding time zone, and the success / failure of the acquired delivery. If indicates a delivery failure, the pattern information may be extracted by setting the time zone (after the fact) after the success / failure time corresponding to the delivery as the surrounding time zone. That is, the pattern information extraction unit 122 may extract the delivery success power pattern from the time zone prior to the success / failure time and extract the delivery failure power pattern from the time zone after the success / failure time. Thereby, the pattern information can be efficiently extracted from each of the time zone having a high probability of including the delivery success power pattern and the time zone having a high probability of including the delivery failure power pattern. However, the embodiment of the present invention is not limited to this, and pattern information may be extracted from both the time zone before and after the success / failure time regardless of the success / failure of delivery.

By registering a characteristic pattern of the change tendency of the power consumption in advance and matching it with the power consumption data of the time zone around the success / failure time, the individual power consumption pattern is extracted and stored in the power consumption pattern DB 113. You may. For example, if the power consumption pattern of a power device can be obtained in advance, or if the use of a specific device is clear from the confirmation of the actual power consumption by the system operator, this is specified as the power consumption pattern of the specific device. It shall be equipped with a function that can be registered as a master. This makes it possible to assist the learning of the system and perform interpolation.

The explanation will be continued by returning to FIG. The pattern information extraction unit 122 collects a plurality of individual power consumption patterns that are similar for each success or failure and stores them in the power consumption pattern DB 113 as a normalized power consumption pattern (S135). That is, the pattern information extraction unit 122 performs pattern matching between individual power consumption patterns, and if they match within a predetermined range, it determines that they are similar. Specifically, the pattern information extraction unit 122 determines that the individual power consumption patterns are similar when the order of the features is the same and the amount of power change for each time interval is within a predetermined range. Then, the pattern information extraction unit 122 adds the number of occurrences of the normalized delivery success power pattern when the similar pattern is the delivery success power pattern. Further, the pattern information extraction unit 122 adds the number of occurrences of the normalized delivery failure power pattern when the similar pattern is the delivery failure power pattern. Then, the pattern information extraction unit 122 calculates the delivery success probability of the normalized delivery success power pattern. Here, it can be calculated by the delivery success probability = the number of delivery success power patterns / (the number of delivery success power patterns + the number of delivery failure power patterns). The pattern information extraction unit 122 may calculate the delivery failure probability of the normalized delivery failure power pattern. Here, it can be calculated by the delivery failure probability = the number of delivery failure power patterns / (the number of delivery success power patterns + the number of delivery failure power patterns). The pattern information extraction unit 122 may not only perform pattern matching between individual power consumption patterns, but may also perform matching between existing normalized power consumption patterns and individual power consumption patterns.

FIG. 10 is a diagram showing an example of a normalized delivery successful power pattern according to the second embodiment of the present invention. FIG. 11 is a diagram showing an example of a normalized delivery failure power pattern according to the second embodiment of the present invention.

FIG. 12 is a diagram for explaining the concept of matching of power consumption patterns according to the second embodiment of the present invention. In the pattern matching of the power consumption pattern, the time interval and the power change amount of the individual power consumption pattern of FIG. 8 and the time interval and the power change amount of the normalized power consumption pattern of FIG. 10 are superimposed for the power consumption pattern to be compared. The power consumption patterns that match with a difference within a certain range can be regarded as similar, that is, the same pattern (the same power equipment is operating). Therefore, the normalized power consumption pattern can also be regarded as the use of a specific power device.

Further, the delivery schedule management server 100 according to the present embodiment increases the number of home rate determination target data by accumulating the delivery record DB 111 and the power consumption DB 112 to be analyzed, and accumulates variations of the power consumption pattern. In addition, the result is fed back to the system as the delivery success and failure results, so that the prediction accuracy can be improved autonomously, that is, the system learns as so-called AI (artificial intelligence).

The explanation will be continued by returning to FIG. Subsequently, the home rate calculation unit 123 calculates the home rate (S140). Specifically, first, the home rate calculation unit 123 acquires the daily power consumption data of a typical period (for example, a month and a day of the week) at a certain delivery destination, and obtains the power consumption data for each time zone. Power consumption pattern Compare with the normalized power consumption pattern in DB113. That is, the home rate calculation unit 123 has the normalized delivery success power pattern and the normalized delivery failure power pattern in the power consumption pattern DB 113 with respect to the time-series power consumption for each consumer acquired from the power consumption DB 112. Pattern match with each. Then, the home rate calculation unit 123 totals the delivery success probability of the normalized power consumption pattern determined to be similar by pattern matching as the home rate in that time zone. At this time, if the occurrence timing of the normalized power consumption pattern is "advance", the home rate of the time zone after the time zone corresponding to the normalized power consumption pattern is delivered (for example, the normalized delivery successful power pattern). Update with success probability. In addition, when the occurrence timing of the normalized power consumption pattern is "after the fact", the home rate of the time zone before the time zone corresponding to the normalized power consumption pattern is determined as the delivery success (for example, the normalized delivery failure power pattern). Update with probability. Here, the home rate is set for each specific time zone of the month, day of the week, and day, and for each delivery destination. That is, the home rate calculation unit 123 performs statistical processing such as averaging the delivery success probability determined to be similar by pattern matching for each specific time zone and delivery destination, and uses the result as the home rate in that time zone. ..

Here, a method of predicting the home rate by a statistical method will be described. First, the home rate calculation unit 123 changes the characteristic power consumption pattern of the daily time-series power consumption shown in FIG. 13 and the normalized power consumption pattern of FIG. 10 or 11 into the pattern shown in FIG. Comparison is performed by a matching method, and if they match, the delivery success probability of the power consumption pattern is totaled for each month × day × time zone, and stored in the home rate DB 114 as the home rate. This tabulation is performed for the entire period of electricity consumption accumulated for each consumer.

FIG. 13 is a diagram showing an example of the power consumption of the home rate calculation target according to the second embodiment of the present invention. As shown in FIG. 13, the characteristic power consumption pattern Pa around 7:00 am of the daily power consumption on XX days of the week matched the normalized power consumption pattern Pb in FIG. As a total, the total number at 7:00 on XX days of the week is multiplied by a certain multiplier to the 86% delivery success probability of the normalized power consumption pattern Pb. In this way, in the analysis of power consumption, by performing pattern matching analysis of the characteristic part of this power consumption pattern, it is considered that the power device is in use (operating), and the home rate during use of the power device is determined. Can be predicted.

FIG. 14 is a diagram showing an example of home rate information of a specific consumer according to the second embodiment of the present invention. As described above, according to the second embodiment of the present invention, it is possible to obtain the home rate at a specific time on a specific day of the specific month for each consumer. When the power consumption for one year is accumulated, the sampling data for obtaining the home rate at a specific time at a certain point is the unit time for the number of days of the week (4 times if there are 4 weeks in a month). It becomes the data of. In an actual system, the aggregation cycle is set by adjusting the accumulation period of power consumption and the aggregation cycle (such as every 3 months instead of monthly) according to the system operation and data accuracy.

The explanation will be continued by returning to FIG. The delivery schedule creation unit 124 creates a delivery schedule with reference to the home rate DB 114 (S150). That is, the delivery schedule creation unit 124 creates a delivery schedule including the delivery order, route, scheduled delivery time, etc., based on the list of delivery destinations scheduled for today's delivery and the home rate of each delivery destination, and the delivery schedule. Save in DB115. Specifically, the delivery schedule creation unit 124 acquires from the home rate DB 114 the time zone in which the home rate is highest on the scheduled delivery day of the consumer's home at the delivery destination in the scheduled delivery of the day, and the scheduled delivery time. And. The delivery company can expect an improvement in the delivery success rate by performing the delivery collectively for each time zone when the home rate is high.

The output unit 125 reads the delivery schedule stored in the delivery schedule DB 115 and transmits it to the mobile terminal of the driver in charge of delivery (S160). Therefore, the delivery schedule, which is information on the delivery route and the scheduled delivery time, which is summarized in the time zone where the home rate is high, can be displayed on the mobile terminal of the driver in charge of delivery, that is, on the map at the delivery site.

As described above, in the embodiment of the present invention, the power consumption pattern in which the consumer is likely to be at home is analyzed, and the time zone in which the consumer is likely to be at home is analyzed from the power consumption status for each consumer. Based on the analysis results, the delivery time and route that can be delivered in a time zone with a high probability of staying at home are calculated, a delivery schedule is created, and the delivery schedule can be referred to from the mobile terminal 200 owned by the delivery person. .. That is, as a method for appropriately adjusting the scheduled delivery time for each consumer's home, attention is paid to changes in power consumption for each consumer, the home rate is analyzed, and delivery efficiency is improved. The delivery success rate can be improved.

In some existing methods, the delivery schedule is sent to the delivery destination user by e-mail in advance, and the delivery schedule can be changed by the user's operation. However, in reality, since it is possible to receive redelivery in the absence of the user, there is little merit for the user, although there is a load on the user's operation. On the other hand, in the embodiment of the present invention, since the success or failure of delivery can be predicted with high accuracy without requiring the operation of the user (delivery destination), the probability of staying at home at an unrecorded delivery destination can be calculated with high accuracy, and redelivery can be performed. You can create a delivery schedule that reduces the probability.

In addition, the embodiment of the present invention can be paraphrased as follows. That is, it is desirable that the pattern information extraction unit determines the characteristics of the change tendency according to the amount of change in power consumption in time series, and extracts the pattern information based on the set of the determined characteristics within a predetermined period. ..

In addition, the prediction unit is at home according to the success or failure of the similar pattern information in the time zone including the change tendency similar to the extracted pattern information in the power consumption information at the arbitrary delivery destination. It is desirable to predict the probability.

Further, the pattern information extraction unit extracts a plurality of change trends in any time zone among the surrounding time zones based on the delivery record information and the power consumption information for each success or failure, and the pattern information extraction unit of each change tendency. Of these, similar ones are grouped and extracted as the pattern information, the degree of success of the delivery among the grouped change trends is calculated as the degree of success of delivery in the pattern information, and the prediction unit performs the comparison. Based on the degree of success of delivery corresponding to the pattern information determined to be similar, the probability of staying at home in the arbitrary time zone may be predicted as the success or failure of the delivery.

Further, the prediction unit determines whether or not the power consumption information at the arbitrary delivery destination is similar to the extracted pattern information, and if it is determined to be similar, the delivery corresponding to the extracted pattern information. It is advisable to predict the success or failure of the delivery by using the degree of success as the probability of staying at home in the time zone when it is determined that they are similar.

Further, when the success / failure time is the success of the delivery, the pattern information extraction unit sets the time zone before the success / failure time as the peripheral time zone, and when the success / failure time is the failure of the delivery. , The pattern information may be extracted by setting the time zone after the success / failure time as the surrounding time zone.

Further, the delivery record storage unit that stores the delivery record information and the success / failure and success / failure time of the actual delivery according to the prediction result are added to the delivery record information at the arbitrary delivery destination, and the delivery record is added. The pattern information extraction unit may further include a storage unit to be stored in the storage unit, and the pattern information extraction unit may read the delivery record information from the delivery record storage unit when extracting the pattern information.

Further, the pattern information extraction unit calculates the amount of power change between the extreme values of power consumption in the surrounding time zone and the time interval between the extreme values, and sets the amount of power change and the time interval. Based on this, the characteristics of the change at the time interval may be determined, and the pattern information may be extracted based on the time-series order of the determined characteristics.

<Embodiment 3>
The third embodiment is a modification of the second embodiment described above. In addition to the second embodiment, the mobile terminal 200 according to the present embodiment can display the latest home-based rate of the scheduled delivery home. Therefore, the pattern information extraction unit according to the present embodiment is based on the power consumption information at the delivery destination on the day in the delivery schedule created based on the result of the prediction, from the scheduled delivery time to a predetermined time before. The change tendency of the power consumption in the latest time zone is compared with the extracted pattern information, and the prediction unit determines that the change tendency in the latest time zone is similar to the extracted pattern information in the pattern information. The prediction of success or failure of delivery in the latest time zone is updated according to the success or failure. As a result, even if the scheduled delivery time is scheduled in advance based on the delivery schedule in a time zone with a high home rate, even if the customer is absent on the delivery date by chance, the power consumption state closest to the scheduled delivery time is reached. It is possible to predict the success or failure of delivery with higher accuracy, and the redelivery rate can be reduced.

Further, it is desirable to further include a notification unit (for example, an output unit 125) that notifies the mobile terminal carried by the delivery person during delivery based on the delivery schedule based on the updated prediction result. As a result, the delivery person can check the latest home-based rate immediately before the delivery, and the delivery can be efficiently performed at the site. Since the configuration of the delivery schedule management server 100 according to the third embodiment is the same as that of FIG. 3, illustration and detailed description thereof will be omitted.

FIG. 15 is a flowchart showing the flow of the latest home rate calculation process according to the third embodiment of the present invention. Here, for example, it is assumed that the scheduled delivery time of a certain delivery destination on the delivery schedule is 15 minutes after the current time. Then, it is assumed that the power consumption data from the current time to 30 minutes before is already stored in the power consumption DB 112. At this time, the pattern information extraction unit 122 acquires the latest power consumption of the delivery destination (S210). Specifically, the pattern information extraction unit 122 performs 1 to 2 hours before the latest acquired time (here, 30 minutes before the current time) of the power consumption data of the delivery destination in the power consumption DB 112. Read the power consumption data for the time period up to.

Next, the pattern information extraction unit 122 makes a comparison with the power consumption pattern (S220). That is, the pattern information extraction unit 122 compares the acquired latest tendency of change in power consumption with each power consumption pattern stored in the power consumption pattern DB 113, and determines whether or not they are similar. FIG. 16 is a diagram showing an example of the latest power consumption according to the third embodiment of the present invention. The power consumption pattern Pc is the power consumption pattern in the latest time zone.

Subsequently, the home rate calculation unit 123 calculates the latest home rate (S230). That is, when the change tendency in the latest time zone is similar to any of the power consumption patterns, the home rate calculation unit 123 updates the home rate in the latest time zone based on the delivery success probability in the power consumption pattern. .. For example, when the power consumption pattern Pc of FIG. 16 is similar to the normalized power consumption pattern Pb of FIG. 10, the home rate calculation unit 123 sets the current home rate of the delivery destination to the successful delivery of the normalized power consumption pattern Pb. Update with a probability of 86%. If the latest power consumption pattern Pc is similar to a plurality of normalized power consumption patterns, the current home rate is updated by multiplying each of them at a fixed ratio.

After that, the output unit 125 notifies the mobile terminal 200 of the latest home rate (S240). As a result, the mobile terminal 200 displays the notified latest home-based rate on the screen, and the delivery person can grasp the latest home-based rate at the scheduled delivery time at the delivery destination. Therefore, it is possible to postpone delivery to a consumer's home with a high absentee rate, prioritize delivery to a consumer's home with a high home-based rate, and improve the delivery success rate by flexibly responding at the site.

<Other embodiments>
In the above embodiment, the home rate for each individual delivery destination is calculated to predict the success or failure of delivery, but the present invention is not limited to these, and the power consumption at a plurality of delivery destinations having some common attribute is used. The consumption pattern may be obtained and the success or failure of general-purpose delivery may be predicted for each attribute. For example, the age group, gender, number of residents, area, etc. of the resident of the delivery destination may be used as attributes. Furthermore, even if the number of residents is one, it is desirable to separate the attributes because the pattern of time spent at home differs depending on the elderly, single employees, students, etc. In addition, it is desirable to separate the attributes because the power consumption tendency differs depending on the area, such as the city center or region, and the northern or southern countries. Then, these attributes may be added to the power consumption or the delivery record.

The present invention is not limited to the above embodiment, and can be appropriately modified without departing from the spirit.

Further, in the above-described embodiment, the present invention has been described as a hardware configuration, but the present invention is not limited thereto. The present invention can also be realized by causing a CPU (Central Processing Unit) to execute a computer program.

In the above example, the program can be stored and supplied to the computer using various types of non-transitory computer readable medium. Non-temporary computer-readable media include various types of tangible storage media. Examples of non-temporary computer-readable media include magnetic recording media (eg, flexible discs, magnetic tapes, hard disk drives), magneto-optical recording media (eg, magneto-optical discs), CD-ROMs (Read Only Memory), CD-Rs. CD-R / W, DVD (Digital Versatile Disc), BD (Blu-ray (registered trademark) Disc), semiconductor memory (for example, mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, RAM (for example) Random Access Memory)) is included. The program may also be supplied to the computer by various types of transient computer readable medium. Examples of temporary computer-readable media include electrical, optical, and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire and an optical fiber, or a wireless communication path.

Some or all of the above embodiments may also be described, but not limited to:
(Appendix 1)
Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is A pattern information extraction unit that extracts pattern information indicating a change tendency for each success or failure,
A prediction unit that compares the power consumption information at an arbitrary delivery destination with the extracted pattern information and predicts the success or failure of delivery for each time zone at the arbitrary delivery destination.
Information processing device equipped with.
(Appendix 2)
The pattern information extraction unit
When the pattern information is extracted in a time zone before the success / failure time, it is used as the pattern information in advance of the delivery, and when the pattern information is extracted in the time zone after the success / failure time, it is after the delivery. The information processing apparatus according to Appendix 1, which is extracted as pattern information.
(Appendix 3)
The prediction unit
Of the power consumption information at the arbitrary delivery destination, for a time zone including a change tendency similar to the extracted pattern information, the home-based probability according to the success or failure of the similar pattern information is predicted. The information processing apparatus according to 2.
(Appendix 4)
The pattern information extraction unit
Based on the delivery record information and the power consumption information, a plurality of change trends in any time zone among the surrounding time zones are extracted for each success or failure.
Similar ones of each change tendency are grouped and extracted as the pattern information.
Of the grouped change trends, the degree of success of the delivery is calculated as the degree of success of delivery in the pattern information.
The prediction unit
The information processing apparatus according to Appendix 1 or 2, which predicts the probability of being at home in the arbitrary time zone as the success or failure of the delivery based on the delivery success degree corresponding to the pattern information determined to be similar by the comparison.
(Appendix 5)
The prediction unit
From the power consumption information at the arbitrary delivery destination, it is determined whether or not the pattern information is similar to the extracted pattern information.
When it is determined that they are similar, the degree of success of delivery corresponding to the extracted pattern information is used as the probability of staying at home in the time zone when it is determined that they are similar. The information processing device described.
(Appendix 6)
The pattern information extraction unit
Based on the power consumption information at the delivery destination of the day in the delivery schedule created based on the result of the prediction, the change tendency of the power consumption in the latest time zone from the scheduled delivery time to the predetermined time before, and the above-extracted Compare with pattern information,
The prediction unit
If the change tendency in the latest time zone is similar to the extracted pattern information, the prediction of success or failure of delivery in the latest time zone is updated according to the success or failure of the pattern information. The information processing apparatus according to item 1.
(Appendix 7)
A delivery record storage unit that stores the delivery record information and
A storage unit that adds the success / failure time and success / failure time of the actual delivery according to the result of the prediction to the delivery record information at the arbitrary delivery destination and stores the delivery record storage unit is further provided.
The information processing device according to any one of Supplementary note 1 to 6, wherein the pattern information extraction unit reads the delivery record information from the delivery record storage unit when extracting the pattern information.
(Appendix 8)
The pattern information extraction unit
Judge the characteristics of the change tendency according to the amount of change in power consumption over time,
The information processing apparatus according to any one of Supplementary note 1 to 7, which extracts the pattern information based on the set of the determined features within a predetermined period.
(Appendix 9)
Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is Pattern information showing a change tendency is extracted for each success or failure, and
By comparing the power consumption information at an arbitrary delivery destination with the extracted pattern information, the success or failure of delivery for each time zone at the arbitrary delivery destination is predicted.
A method of predicting achievement or failure of distribution.
(Appendix 10)
Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is The process of extracting pattern information indicating a change tendency for each success or failure, and
A process of comparing the power consumption information at an arbitrary delivery destination with the extracted pattern information to predict the success or failure of delivery for each time zone at the arbitrary delivery destination.
Achievement / failure prediction program that causes a computer to execute.
(Appendix 11)
The pattern information extraction unit
The amount of power change between the extreme values of power consumption in the surrounding time zone and the time interval between the extreme values are calculated.
Based on the power change amount and the time interval, the characteristics of the change in the time interval are determined.
The information processing apparatus according to Appendix 8, which is extracted as the pattern information based on the time-series order of the determined features.
(Appendix 12)
The information processing apparatus according to Appendix 6, further comprising a notification unit that notifies a mobile terminal carried by a delivery person who is delivering based on the delivery schedule based on the updated prediction result.
(Appendix 13)
The pattern information extraction unit
When the success or failure of the delivery indicates the success of the delivery, the time zone before the success or failure time corresponding to the delivery is set as the surrounding time zone, and when the success or failure of the delivery indicates the failure of the delivery, the delivery is supported. The information processing apparatus according to Appendix 1, wherein the pattern information is extracted with a time zone after the success / failure time as the peripheral time zone.

1 Information processing device 11 Pattern information extraction unit 12 Prediction unit 21 Delivery record information 22 Power consumption information 100 Delivery schedule management server 111 Delivery record DB
112 Power consumption DB
113 Power consumption pattern DB
114 Home rate DB
115 Delivery schedule DB
116 Delivery destination list DB scheduled for delivery today
121 Collection unit 122 Pattern information extraction unit 123 Home rate calculation unit 124 Delivery schedule creation unit 125 Output unit 200 Mobile terminal PP1 Power consumption pattern PP2 Power consumption pattern PP3 Power consumption pattern Pa Power consumption pattern Pb Normalized power consumption pattern Pc Power consumption pattern

Claims (1)

Based on the delivery record information including the success / failure and success / failure time of delivery to a predetermined delivery destination and the power consumption information indicating the time-series power consumption at the delivery destination, the power consumption in the time zone around the success / failure time is Pattern information showing a change tendency is extracted for each success or failure, and
The characteristics of the change tendency are determined according to the amount of change in power consumption over time, and the pattern information is extracted based on the set of the determined characteristics within a predetermined period.
The amount of power change between the extreme values of power consumption in the surrounding time zone and the time interval between the extreme values are calculated, and the change in the time interval is based on the amount of power change and the time interval. A pattern information extraction unit that determines the characteristics of the above-mentioned features and extracts them as the pattern information based on the time-series order of the determined features.
A prediction unit that compares the power consumption information at an arbitrary delivery destination with the extracted pattern information and predicts the success or failure of delivery for each time zone at the arbitrary delivery destination.
Information processing device.

Images