Calculation device, calculation method, and calculation program

Abstract

This calculation device includes a calculation unit that calculates prediction data representing, for each predetermined time unit, the number of users expected to visit each of a plurality of visit destinations on the basis of performance data representing the results of the number of users having visited each of a plurality of predetermined visit destinations for each predetermined time unit, a past visit plan generated for each user and representing for each time unit a visit destination to be recommended to the user, as well as corresponding to the performance data, and a future visit plan from the current time point and which is the visit plan generated for each user.

Description

Calculation device, calculation method, and calculation program

[0001] The technology of the present disclosure relates to a calculation device, a calculation method, and a calculation program.

[0002] A technology is known that dynamically recommends tour schedules to ensure that all customers can move around comfortably in facilities such as tourist spots and large-scale event venues by making spatiotemporal predictions of near-future congestion conditions or resource demands based on real-time observations of people and traffic flows and search logs of desired attractions or destinations (Non-Patent Document 1).

[0003] "Dynamic Tour Schedule Recommendation Using Spatiotemporal Prediction and Mathematical Optimization Technologies," NTT Communication Science Laboratories, Internet Search <URL: https: / / www.rd.ntt / research / CS0024.html>

[0004] In the prior art, when generating a visiting plan such as a tour schedule, congestion is taken into consideration, but the congestion that is taken into consideration is the actual number of people, and future congestion is not taken into consideration.

[0005] For example, if a user creates a visit plan that avoids crowds and follows it, the congestion situation will likely be different from past visits. However, this cannot be taken into account if only the actual number of people is looked at. The same destination may be recommended to many people, and the destination may end up being crowded with people trying to avoid the crowds.

[0006] On the other hand, users do not necessarily follow recommended visit plans, and visit plans are not necessarily generated for all users, so visit plans cannot simply be thought of as future congestion.

[0007] The disclosed technology has been developed in consideration of the above points, and aims to provide a calculation device, calculation method, and calculation program that can calculate the expected number of users at each destination, taking into account that users may not necessarily visit according to the recommended visiting plan.

[0008] A first aspect of the present disclosure is a calculation device that includes: performance data representing the actual number of users who have visited each of a predetermined number of destinations for each predetermined time unit; a visiting plan generated for each user representing destinations recommended to the user for each time unit, the visiting plan corresponding to the performance data; and the visiting plan generated for each user representing a future time period beyond the present time, the calculation unit calculating forecast data representing the number of users expected to visit each of the plurality of destinations for each predetermined time unit.

[0009] A second aspect of the present disclosure is a calculation method in which a computer calculates forecast data representing the number of users expected to visit each of a plurality of destinations for each predetermined time unit, based on performance data representing the actual number of users who have visited each of a plurality of predetermined destinations for each predetermined time unit, a visiting plan generated for each user representing destinations recommended to the user for each time unit, the visiting plan corresponding to the performance data, and the visiting plan generated for each user for a future time period beyond the present.

[0010] A third aspect of the present disclosure is a calculation program that causes a computer to calculate forecast data representing the number of users expected to visit each of a plurality of destinations for each predetermined time unit, based on performance data representing the actual number of users who have visited each of a plurality of predetermined destinations for each predetermined time unit, a visiting plan generated for each user representing destinations recommended to the user for each time unit, the visiting plan corresponding to the performance data, and the visiting plan generated for each user for a future time period beyond the present.

[0011] According to the disclosed technology, it is possible to calculate the predicted number of users at each destination, taking into consideration that users may not necessarily visit according to the recommended visiting plan.

[0012] 1 is a schematic block diagram of an example of a computer functioning as a calculation device of this embodiment. FIG. 1 is a block diagram showing the functional configuration of the calculation device of this embodiment. FIG. 2 is a diagram showing an example of performance data. FIG. 3 is a diagram showing an example of a visiting plan. FIG. 4 is a block diagram showing the functional configuration of a calculation unit of the calculation device of this embodiment. FIG. 5 is a diagram showing an example of the number of recommended users. FIG. 6 is a diagram showing a visiting plan generated for each user and an example of the number of users recommended to visit each destination for each period. FIG. 7 is a diagram showing an example of the number of users recommended in the past. FIG. 8 is a diagram showing an example of the number of users recommended in the past to visit each destination for each period. FIG. 9 is a diagram showing an example of the performance of the number of users who have visited each destination for each period. FIG. 10 is a diagram showing an example of a ratio calculated for each destination and for each period. FIG. 11 is a diagram showing an example of a product calculated for each destination and for each period. FIG. 12 is a diagram showing an example of a ratio of the product for each period to the sum of the products for each period. FIG. 13 is a diagram showing an example of predicted data for the number of users. FIG. 14 is a block diagram showing the functional configuration of a plan generation unit of the calculation device of this embodiment. FIG. 15 is a diagram showing an example of setting an entry time, an exit time, a lunch time, a dinner time, and a shopping time. FIG. 16 is an example of a recommendation score for each destination. FIG. 17 is an example of a crowd avoidance score for each destination for each period. 1 is an example of a selection probability for each destination and for each time unit. FIG. 2 is a diagram showing an example of weighting for each of congestion avoidance, popularity priority, number of visits priority, and destination preference priority. FIG. 3 is a diagram showing an example of weighting for each of destinations, going home, lunch, dinner, and shopping. FIG. 4 is a diagram showing an example of weighting for each destination. FIG. 5 is a diagram showing an example of weighting for each destination. FIG. 6 is a flowchart showing the flow of a plan generation process of this embodiment. FIG. 7 is a flowchart showing the flow of a process for calculating forecast data in the plan generation process of this embodiment. FIG. 8 is a flowchart showing the flow of a process for generating a future visiting plan in the plan generation process of this embodiment.

[0013] An example of an embodiment of the disclosed technology will be described below with reference to the drawings. Note that the same reference numerals are used to designate identical or equivalent components and parts in each drawing. Also, the dimensional proportions in the drawings are exaggerated for the sake of explanation and may differ from the actual proportions.

[0014] <Outline of this embodiment> In this embodiment, a situation is considered in which multiple users take turns visiting multiple destinations of their own volition. A visiting plan is recommended to the users and they are asked to follow it, so that the destinations do not get crowded.

[0015] A visit plan is generated after the user indicates their intention to create a destination. When generating a visit plan, information such as the number of people at a particular destination at a particular time is used and incorporated into the visit plan to generate a visit plan that can avoid congestion. Also, it is assumed that the user does not necessarily follow the generated visit plan.

[0016] In this embodiment, future congestion conditions are predicted while taking into consideration that users may not necessarily follow recommended visit plans. A visit plan that avoids congestion is generated by calculating the ratio of the number of users recommended in past visit plans to the number of users who actually followed the visit plan, and estimating the number of users in the future taking into consideration how many users will follow the visit plan.

[0017] <Configuration of Calculation Device According to This Embodiment> FIG. 1 is a block diagram showing the hardware configuration of a calculation device 10 according to this embodiment.

[0018] 1 , the computing device 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM 13, a storage 14, an input unit 15, a display unit 16, and a communication interface (I / F) 17. Each component is connected to each other via a bus 19 so as to be able to communicate with each other.

[0019] The CPU 11 is a central processing unit that executes various programs and controls each component. That is, the CPU 11 reads programs from the ROM 12 or the storage 14 and executes the programs using the RAM 13 as a work area. The CPU 11 controls the above components and performs various arithmetic processing in accordance with the programs stored in the ROM 12 or the storage 14. In this embodiment, a calculation program is stored in the ROM 12 or the storage 14. The calculation program may be a single program or a group of programs composed of multiple programs or modules.

[0020] The ROM 12 stores various programs and various data. The RAM 13 temporarily stores programs or data as a working area. The storage 14 is configured with an HDD (Hard Disk Drive) or an SSD (Solid State Drive) and stores various programs including the operating system and various data.

[0021] The input unit 15 includes a pointing device such as a mouse and a keyboard, and is used to input various types of information.

[0022] The display unit 16 is, for example, a liquid crystal display, and displays various information including the processing results. The display unit 16 may be a touch panel type and function as the input unit 15.

[0023] The communication interface 17 is an interface for communicating with other devices, and uses standards such as Ethernet (registered trademark), FDDI, and Wi-Fi (registered trademark).

[0024] Next, a description will be given of the functional configuration of the calculation device 10. Fig. 2 is a block diagram showing an example of the functional configuration of the calculation device 10.

[0025] As shown in FIG. 2, the calculation device 10 functionally comprises a performance database (DB) 20, a visit plan database (DB) 22, a calculation unit 24, and a plan generation unit 26.

[0026] The performance database (DB) 20 stores performance data that indicates the performance of the number of users who visited each of a plurality of predetermined destinations for each predetermined time period.

[0027] For example, as shown in FIG. 3 , the performance database 20 stores performance data indicating the performance of the number of users for each date and time for each visited destination. FIG. 3 shows an example of performance data indicating that the number of users is "0" at the date and time "2023 / 10 / 23 0:00" for the visited destination ID "1." The contents of the performance database 20 are updated according to the performance. The performance database 20 may always store the latest performance data, or the performance data may be updated every hour. In this embodiment, an example in which one hour is used as an example of a time unit will be described, but the time unit is not limited to this, and minutes, days, months, etc. may also be used as the time unit.

[0028] The visiting plan database (DB) 22 stores past visiting plans that are generated for each user and represent visiting destinations recommended to the user for each period of time. The visiting plan database (DB) 22 also stores future visiting plans that are generated for each user and extend beyond the present time.

[0029] A visiting plan is data in which a visiting destination ID is linked to each time period for each user and each date, as shown in Fig. 4. The example in Fig. 4 indicates that a visit will be made to a visiting destination with a visiting destination ID of "3" from 10:00 to 12:00.

[0030] The calculation unit 24 calculates forecast data representing the number of users who will visit each of a plurality of destinations for each predetermined time period, based on the performance data stored in the performance database 20 and the past visit plans and future visit plans corresponding to the performance data, which are stored in the visit plan database 22.

[0031] Specifically, as shown in FIG. 5, the calculation unit 24 includes a future plan acquisition unit 30, a user number calculation unit 32, a past plan acquisition unit 34, a user number calculation unit 36, a performance data acquisition unit 38, a ratio calculation unit 40, a product calculation unit 42, a percentage calculation unit 44, a sum calculation unit 46, and a forecast data calculation unit 48.

[0032] The future plan acquisition unit 30 acquires visiting plans for the future from the present time that are stored in the visiting plan database 22. Specifically, it acquires visiting plans created for future dates. By using visiting plans for the future from the present time, it is possible to calculate trends in places and times where people are likely to congregate, and the visiting plans created by the plan creation unit 26 are able to avoid congestion.

[0033] The user number calculation unit 32 calculates, for each destination and for each time period, the number of users recommended to visit the destination for each time period in the future visit plan, based on the visit plans from the present time onward acquired by the future plan acquisition unit 30 (see FIG. 6 ). FIG. 6 shows an example in which the number of users recommended to visit destination ID "1" at time "14:00" is "0," the number of users recommended to visit destination ID "1" at time "15:00" is "80," and the number of users recommended to visit destination ID "1" at time "16:00" is "40."

[0034] For example, as shown in FIG. 7 , assume that a visiting plan is generated for user 1 between 10:00 and 11:00, and visiting plans are generated for user 2 and user 3 between 11:00 and 12:00. In this case, since no visiting plan has been generated for anyone at 10:00, the number of users recommended to visit destination ID "1" in each time unit is calculated to be "0." Since a visiting plan has been generated for user 1 at 11:00, the number of users recommended to visit destination ID "1" at 13:00 is calculated to be "1." Since visiting plans have been generated for user 1, user 2, and user 3 at 12:00, the number of users recommended to visit destination ID "1" before 13:00 is calculated to be "2."

[0035] The past plan acquisition unit 34 acquires past visiting plans stored in the visiting plan database 22. Specifically, it acquires visiting plans created for past dates.

[0036] The user number calculation unit 36 ​​calculates, for each destination and for each period, the number of users who were recommended to visit the destination in the past visit plan for that period, based on the past visit plans acquired by the past plan acquisition unit 34 (see FIG. 8 ). FIG. 8 shows an example in which the number of users who were recommended to visit destination ID "1" at time "14:00" is "50," the number of users who were recommended to visit destination ID "1" at time "15:00" is "70," and the number of users who were recommended to visit destination ID "1" at time "16:00" is "90."

[0037] As shown in Figure 9, past visit plans can reveal trends in which destinations are likely to be recommended at what times. It is recommended to use past visit plans generated for as many dates as possible, as this makes it easier to read trends. In this case, the trend in the number of users can be read by calculating the average number of users for each period from the past visit plans generated for multiple dates.

[0038] The performance data acquiring unit 38 acquires, for each destination and for each period, the performance of the number of users who visited the destination in that period (see FIG. 10).

[0039] FIG. 10 shows an example in which the actual number of users who visited destination ID "1" at time "14:00" is "500," the actual number of users who visited destination ID "1" at time "15:00" is "350," and the actual number of users who visited destination ID "1" at time "16:00" is "900."

[0040] The ratio calculation unit 40 calculates, for each destination and each time period, the ratio between the number of users who were recommended to visit the destination in the past visit plan during that time period and the actual number of users who visited the destination during that time period.

[0041] Specifically, the ratio is calculated by dividing the actual number of users by the number of recommended users for each visit destination ID and each period (see FIG. 11 ). The number of users recommended in past visit plans used to calculate the ratio may be, for example, an average of the number of users over multiple days. The actual number of users used to calculate the ratio may also be, for example, an average of the actual number of users over multiple days.

[0042] FIG. 11 shows an example in which the ratio for the time "14:00" and the visiting destination ID "1" is "10" (=500 / 50), the ratio for the time "15:00" and the visiting destination ID "1" is "5" (=350 / 70), and the ratio for the time "16:00" and the visiting destination ID "1" is "10" (=900 / 90). This ratio takes into consideration how likely a user is to follow the recommended visiting plan. Because the ratio is calculated for each visiting destination ID and for each time period, it is possible to reflect differences in the ease of following the visiting plan depending on the visiting destination and time.

[0043] The product calculation unit 42 calculates the product of the number of users recommended to visit the destination in the time period in the future visit plan for each destination and for each time period by the calculated ratio ( FIG. 12 ). Specifically, the ratio for each destination ID and for each time period calculated by the ratio calculation unit 40 is multiplied by the number of users for each destination ID and for each time period calculated by the user number calculation unit 32, and the product is calculated as the number of recommended users taking into account the actual number of users. FIG. 12 shows an example in which the product of the time "14:00" and the destination ID "1" is "0" (=0×10), the product of the time "15:00" and the destination ID "1" is "400" (=80×5), and the product of the time "16:00" and the destination ID "1" is "400" (=40×10).

[0044] The ratio calculation unit 44 calculates, for each destination, the ratio of the product for each period to the sum of the products for each period ( FIG. 13 ). In FIG. 13 , for destination ID “1,” the ratio for the time “14:00” is “0” (=0 / 800), the ratio for the time “15:00” is “0.5” (=400 / 800), and the ratio for the time “16:00” is “0.5” (=400 / 800).

[0045] The sum calculation unit 46 calculates, for each destination, the sum of the number of users who visited that destination, which is obtained from the performance data. For example, for each destination, the sum of the performance numbers of users who visited that destination in one day for each period acquired by the performance data acquisition unit 38 is calculated. In the example of FIG. 10 above, the sum of the performance numbers of users who visited destination ID "1" in one day is calculated to be "1750" (=500+350+900).

[0046] The predicted data calculation unit 48 calculates predicted data for the number of users based on the calculated ratio for each destination and for each time unit and the total number of users who visited the destination obtained from the performance data. Specifically, the predicted data for the number of users is calculated as the product of the calculated ratio for each destination and for each time unit and the total number of users who visited the destination obtained from the performance data ( FIG. 14 ).

[0047] FIG. 14 shows an example in which, for a destination ID of "1," the predicted data for the number of users at the time "14:00" is "0" (=0×1750), the predicted data for the number of users at the time "15:00" is "875" (=0.5×1750), and the predicted data for the number of users at the time "16:00" is "875" (=0.5×1750).

[0048] In this way, the number of predicted data items can be calculated taking into account the ease of following the visiting plan, which may differ depending on the visiting destination and time. However, if there is a visiting destination not included in the visiting plan acquired by the future plan acquisition unit 30, or if there is time available, even if the number of users at that visiting destination or time is multiplied by the ratio, the number of users may remain small, and the overall number of users may end up being extremely small. Therefore, in this embodiment, the predicted data calculation unit 48 can adjust the number of people by allocating the number of users calculated by the sum calculation unit 46 to the ratio calculated by the ratio calculation unit 44.

[0049] In the above example, only the visit ID "1" is listed, but if there are multiple visit IDs, the same process can be performed for each visit ID. Also, although only the period from 2:00 PM to 4:00 PM is listed, this is merely an example, and the same process can be applied to other times.

[0050] The plan generation unit 26 generates a future visiting plan for the user based on the prediction data and the recommendation score obtained for each visiting destination, so as to avoid crowds.

[0051] Specifically, as shown in FIG. 15, the plan generating unit 26 includes a setting unit 50 , a score calculating unit 52 , a selection probability calculating unit 54 , a visiting plan generating unit 56 , and an evaluation unit 58 .

[0052] The setting unit 50 sets the entrance time, exit time, lunch time, dinner time, and shopping time based on input from the user.

[0053] Specifically, as shown in FIG. 16 , the entry time and exit time are determined, and the corresponding sections of the visit plan are filled in. Entering an appointment before the entry time or after the exit time results in an unfeasible visit plan, so such visit plans are filled in first to avoid generation. This reduces the generation of meaningless visit plans and is expected to shorten processing time compared to not filling in first. Furthermore, since lunch, dinner, and shopping are generally considered to occur between the entry time and the exit time, these times are filled in after the entry time and the exit time. It is also assumed that there are desirable time slots for lunch and dinner. It is generally assumed that many people eat lunch around 12:00, and that many people eat dinner after 17:00. By filling in lunch and dinner before the destinations, it is possible to avoid situations where lunch and dinner plans cannot be made because the destinations are scheduled at times suitable for lunch or dinner. It is also assumed that there is an optimal time for shopping. For example, shopping just before leaving can avoid walking around with luggage. By filling in the shopping before the destination, it is possible to avoid situations where a visit to a destination is scheduled at a time that is suitable for shopping and the shopping cannot be scheduled. It also makes it easier to generate a visit plan that matches human sensibilities, which is expected to reduce processing time compared to generating a visit plan that does not match human sensibilities in a wasteful manner.

[0054] In the example shown in Figure 16 above, the time "10:00" is set as the entry time, the time "14:00" is set as the lunch time, the time "18:00" is set as the dinner time, the time "20:00" is set as the shopping time, and the time "21:00" is set as the exit time.

[0055] The score calculation unit 52 calculates a recommendation score for each destination ( FIG. 17 ). A conventionally known method may be used to calculate the recommendation score. FIG. 17 shows an example in which the recommendation score for destination ID “1” is “0.1,” the recommendation score for destination ID “2” is “0.4,” and the recommendation score for destination ID “3” is “0.8.”

[0056] Furthermore, the score calculation unit 52 calculates the congestion avoidance score for each destination for each time period based on the predicted data ( FIG. 18 ). As an example of the calculation method, for each destination and each time period, the predicted data on the number of users is divided by a predetermined maximum number of people (e.g., the number of people that can be accommodated) for that destination to calculate the congestion degree, and then the congestion avoidance score is calculated by subtracting the congestion degree from 1. FIG. 18 shows an example in which the congestion avoidance score for the destination ID "1" at the time "10:00" is "0.6" and the congestion avoidance score at the time "11:00" is "0.4."

[0057] The selection probability calculation unit 54 calculates the selection probability for each destination and each time period based on the recommendation score of each destination and the congestion avoidance score for each destination and each time period (FIG. 19).

[0058] Specifically, the score is calculated by the product, sum, average, or weighted sum of the recommendation score and the congestion avoidance score. ａ，ｈ Calculate the following selection probability ａ，ｈ where a indicates the destination ID and h indicates the time.

[0059]

[0060] FIG. 19 shows an example in which the selection probability of the time "10 o'clock" for the visited destination ID "1" is "0.225" and the selection probability of the time "11 o'clock" is "0.175".

[0061] This selection probability is calculated using both the recommendation score and the congestion avoidance score, making it easier to include many destinations with good recommendation and congestion avoidance scores in your visit plan.

[0062] The visiting plan generation unit 56 determines the visiting destinations for each time period in accordance with the selection probability of each visiting destination in that time period, and applies the determined destinations to the visiting plan, thereby generating a visiting plan (see FIG. 16 above).

[0063] Here, at lunchtime, dinnertime, and shopping time, destinations corresponding to the respective times may be determined using the selection probability. In this case, a flag indicating an action such as "lunch," "dinner," or "shopping" may be attached to the corresponding destination, and destinations at lunchtime, dinnertime, or shopping time may be determined from the destinations attached with the flag according to the selection probability.

[0064] In the example shown in FIG. 16, destination ID "1" is determined as the destination for the times "12:00" and "13:00", destination ID "3" is determined as the destination for the time "16:00", and destination ID "2" is determined as the destination for the time "19:00".

[0065] The visiting plan generation unit 56 generates multiple visiting plans by repeatedly generating visiting plans. By generating multiple visiting plans and selecting from them, it is expected that a good visiting plan can be obtained. For example, when visiting two destinations, if the congestion levels at each destination differ depending on the time period, it is thought that the congestion level of the entire visiting plan will change depending on the order in which the destinations are visited. By generating many visiting plans, it is possible to generate a visiting plan in either order, and it is expected that it will be possible to take into account the fact that switching the time periods will reduce overall congestion. Here, it is necessary to define what a good visiting plan is.

[0066] In this embodiment, the evaluation unit 58 calculates a visit plan evaluation value as a measure of the quality of the visit plan. The visit plan evaluation value is a value that represents the quality of the visit plan as a whole, rather than of each individual destination. The score calculation unit 52 calculates a recommendation score for each destination, but simply packing the top destinations into a visit plan may not result in a good visit plan. For example, if all the destinations with high recommendation scores are likely to be crowded, the visitor will end up being stuck in crowded places all day. Furthermore, by packing the visit plan with destinations that rank high in the congestion avoidance score calculated by the score calculation unit 52, it is expected that the visit plan will be one that makes it easier to avoid congestion, but there is a possibility that popular destinations will not be included in the visit plan. Therefore, if the visit plan is good as a whole, it is designed so that the visit plan evaluation value is high.

[0067] Specifically, the evaluation unit 58 changes the factors that are emphasized for each of the visit plans, namely, avoiding crowds, prioritizing popularity, prioritizing number of visits, and prioritizing destination preferences, and calculates a visit plan evaluation value for each visit plan for whichever of the visit plans selected by the user, namely, avoiding crowds, prioritizing popularity, prioritizing number of visits, and prioritizing destination preferences.

[0068] For example, the visit plan evaluation value for visit plan x is calculated according to the following formula: ｂ (x) is calculated.

[0069] Visit plan evaluation value ｂ (x) = s ｂ，１ × Avoid congestion (x) + s ｂ，２ × Popularity (x) + s ｂ，３ × Number of visits (x) + s ｂ，４ ×Destination preference (x)

[0070] s ｂ，１ , s ｂ，２ , s ｂ，３ , s ｂ，４ is a weight, which is set in advance as shown in FIG. 20 for each of congestion avoidance, popularity priority, number of visits priority, and destination preference priority. ｂ，１ , s ｂ，２ , s ｂ，３ , s ｂ，４The weights may be set so that the sum of the weights is 1. In this case, the degree of attention paid to each element can be seen as a ratio, making it easier to interpret.

[0071] Congestion avoidance (x) is calculated by adding the congestion avoidance scores of the destination, going home, lunch, dinner, and shopping included in visiting plan x, as shown in the following formula.

[0072] Congestion avoidance (x) = h １，１ × destination congestion avoidance score + h １，２ × Congestion avoidance score for returning home + h １，３ × Lunchtime crowd avoidance score + h １，４ × Dinner crowd avoidance score + h １，５ × Shopping crowd avoidance score

[0073] If a user visits multiple destinations, the congestion avoidance score for each destination can be calculated as the average of the congestion avoidance scores for each destination. １，１ , h １，２ , h １，３ , h １，４ , h １，５ is a weight, which is set in advance as shown in FIG. 21. The corresponding h １，３ , h １，４ , h １，５ Set to 0.

[0074] In addition, h １，１ , h １，２ , h １，３ , h １，４ , h １，５ The weights may be set so that the sum of the weights is 1. In this case, the importance of each element can be seen as a percentage, making it easier to interpret. By changing the weight for each element, it is possible to reflect differences in importance, such as avoiding crowds at destinations being preferable to avoiding crowds when shopping.

[0075] The number of visits (x) is calculated by adding the visit number scores of the destinations, lunch, dinner, and shopping included in the visit plan x, as shown in the following formula.

[0076] Number of visits (x) = h ３，１ × Visitor score of destination + h ３，２× Lunch visit score + h ３，３ × Dinner visit score + h ３，４ × Shopping visit score

[0077] If you visit multiple destinations, you can calculate the visit score for each destination as the average of the visit score for each destination. ３，１ , h ３，２ , h ３，３ , h ３，４ is a weight, which is set in advance as shown in FIG. １，１ , h １，２ , h １，３ , h １，４ , h １，５ The weights may be set so that the sum of the weights is 1. In this case, the importance of each element can be seen as a percentage, making it easier to interpret.

[0078] The visit score is calculated using the following formula:

[0079] Visit score = Base visit score - Time required / Visit score adjustment parameter

[0080] If the two items on the left side of the formula are greater than the standard score for the number of visits, the two items on the left side of the formula are used as the standard score for the number of visits.

[0081] The reference visit count score is a predetermined value, for example, set to 1.5. The visit count score adjustment parameter is a predetermined value, for example, set to 60. The required time is a predetermined value for each destination.

[0082] For example, if there is a destination that takes 60 minutes, the visit count score for that destination is calculated as 1.5 - 60 / 60 = 0.5. When the visit count score is calculated as above, it is expected that the visit count score will be a value between 0 and 1, so a destination with a score of 0.5 is considered to be a destination that takes a medium amount of time to visit. In this way, the base visit count score and the visit count score adjustment parameter are adjusted.

[0083] Popularity (x) is calculated by adding the popularity scores of the destinations, lunch, dinner, and shopping included in visiting plan x, as shown in the following formula.

[0084] Popularity (x) = h ２，１ × Popularity score of destination + h ２，２ × popularity score of lunch destination + h ２，３ × popularity score of dinner destination + h ２，４ ×Popularity score of shopping destination

[0085] If you visit multiple destinations, you can calculate the popularity score of each destination as the average of the popularity scores for each destination. ２，１ , h ２，２ , h ２，３ , h ２，４ is a weight, which is set in advance as shown in FIG. ２，１ , h ２，２ , h ２，３ , h ２，４ The weights may be set so that the sum of the weights is 1. In this case, the importance of each element can be seen as a percentage, making it easier to interpret.

[0086] The popularity score of a destination is calculated using the following formula:

[0087] Popularity score = popularity of destination / maximum popularity parameter

[0088] The popularity of a destination is a value that is predetermined for each destination. For example, if the popularity of a destination is 10 and the maximum popularity parameter is set to 20, the popularity score of this destination will be 0.5 (= 10 / 20). If the popularity score of a destination that exceeds the maximum popularity parameter is set to 1, then a destination with a popularity of around 20 (or more) would be considered highly popular, and a destination with a popularity of around 10 would be considered to be about average in popularity.

[0089] Like popularity (x), destination preference (x) is calculated by weighting and adding up the recommendation scores of destinations, lunch destinations, dinner destinations, and shopping destinations included in visiting plan x.

[0090] Visit destination preference (x) = h ４，１ × Destination recommendation score + h ４，２ × Lunch destination recommendation score + h ４，３×Recommendation score for dinner destinations + h ４，４ ×Recommendation score for shopping destinations

[0091] If you are going to multiple destinations, you can calculate the recommendation score for each destination as the average of the recommendation scores for each destination. ４，１ , h ４，２ , h ４，３ , h ４，４ is a weight that is set in advance. ４，１ , h ４，２ , h ４，３ , h ４，４ The weights may be set so that the sum of the weights is 1. In this case, the importance of each element can be seen as a percentage, making it easier to interpret.

[0092] The evaluation unit 58 presents to the user the visiting plan with the highest visiting plan evaluation value from among the generated visiting plans.

[0093] <Operation of Calculation Device According to This Embodiment> Next, the operation of the calculation device 10 according to this embodiment will be described.

[0094] 24 is a flowchart showing the flow of a plan generation process by the calculation device 10. When a plan generation instruction is received from a target user, the CPU 11 reads out a calculation program from the ROM 12 or the storage 14, loads it into the RAM 13, and executes it, thereby performing the plan generation process.

[0095] In step S100, the CPU 11, functioning as the calculation unit 24, calculates forecast data representing the number of users who will visit each of a plurality of destinations for each predetermined time period, based on the performance data stored in the performance database 20 and the past visit plans and future visit plans corresponding to the performance data, which are stored in the visit plan database 22. Here, a past visit plan corresponding to the performance data is, for example, a past visit plan that has the same month, date, or day of the week as the performance data.

[0096] In step S102, the CPU 11, as the plan generation unit 26, generates a future visit plan for the target user based on the forecast data so as to avoid congestion, stores the plan in the visit plan database 22, and presents it to the target user, thereby completing the plan generation process.

[0097] The above step S100 is realized by the processing routine shown in FIG.

[0098] In step S110 , the CPU 11 functions as the future plan acquisition unit 30 to acquire a visiting plan from the present time forward, which is stored in the visiting plan database 22 .

[0099] In step S112, the CPU 11, as the user number calculation unit 32, calculates, for each destination and for each time period, the number of users who are recommended to visit the destination in the future visit plan for that time period based on the visit plan for the future from the current time obtained in step S110.

[0100] In step S114 , the CPU 11 functions as the past plan acquisition unit 34 to acquire past visiting plans stored in the visiting plan database 22 .

[0101] In step S116, the CPU 11, as the user number calculation unit 36, calculates, based on the past visit plans acquired in step S114, the number of users who were recommended to visit the destination in that time period in the past visit plans, for each destination and for each time period.

[0102] In step S118, the CPU 11 functions as the performance data acquisition unit 38 to acquire, for each destination and for each period, the performance of the number of users who visited the destination in that period.

[0103] In step S120, the CPU 11, as the ratio calculation unit 40, calculates, for each destination and each time period, the ratio between the number of users who were recommended to visit the destination in that time period in past visit plans and the actual number of users who visited the destination in that time period.

[0104] In step S122, the CPU 11, as the product calculation unit 42, calculates, for each destination and each time period, the product of the number of users recommended to visit the destination in the future visit plan for that time period and the calculated ratio.

[0105] In step S124, the CPU 11 functions as the ratio calculation unit 44 to calculate, for each destination, the ratio of the product for each period to the sum of the products for each period.

[0106] In step S126, the CPU 11 functions as the sum calculation unit 46 to calculate, for each destination, the sum of the number of users who visited that destination in one day, which is obtained from the performance data.

[0107] In step S128, the CPU 11, as the predicted data calculation unit 48, calculates predicted data for the number of users based on the calculated ratio for each destination and each time unit and the total number of users who visited the destination obtained from the actual data.

[0108] The above step S102 is realized by the processing routine shown in FIG.

[0109] In step S130, the CPU 11 functions as the setting unit 50 to set the entry time and exit time based on input from the user.

[0110] In step S132, the CPU 11 functions as the setting unit 50 to set lunch time, dinner time, and shopping time based on input from the user.

[0111] In step S134, the CPU 11 functions as the score calculation unit 52 to calculate a recommendation score for each destination.

[0112] In step S136, the CPU 11 functions as the score calculation unit 52 to calculate the congestion avoidance score for each destination for each period based on the prediction data.

[0113] In step S138, the CPU 11, as the selection probability calculation unit 54, calculates the selection probability for each destination and each time period based on the recommendation score of each destination and the congestion avoidance score for each destination and each time period.

[0114] In step S140, the CPU 11, as the visiting plan generation unit 56, determines the visiting destinations for each time period in accordance with the selection probability of each visiting destination in that time period and applies the determined destinations to the visiting plan, thereby generating a visiting plan. The CPU 11, as the visiting plan generation unit 56, generates a plurality of visiting plans by repeating the process of generating visiting plans.

[0115] In step S142, the CPU 11, functioning as the evaluation unit 58, calculates a visiting plan evaluation value for each of the generated visiting plans as an indication of the quality of the visiting plan. At this time, the visiting plan evaluation value is calculated using weights sb,1, sb,2, sb,3, and sb,4 that are determined for one of congestion avoidance, popularity priority, number of visits priority, and destination preference priority, selected by the target user. The CPU 11, functioning as the evaluation unit 58, presents to the user the visiting plan with the highest visiting plan evaluation value out of the generated visiting plans, and stores it in the visiting plan database 22.

[0116] As described above, the calculation device according to this embodiment calculates forecast data representing the number of users expected to visit each of a plurality of destinations for each predetermined time period, based on performance data representing the actual number of users who have visited each of a plurality of destinations for each predetermined time period, visit plans generated for each user representing destinations recommended to the user for each time period, past visit plans corresponding to the performance data, and visit plans generated for each user representing future visit plans from the present time, and generates a future visit plan for the target user based on the forecast data. This makes it possible to calculate the forecast number of users at each destination, taking into account the fact that users may not necessarily visit according to the recommended visit plan. Furthermore, a visit plan to be recommended to the user can be generated based on the forecast number of users at each destination.

[0117] In addition, by using future visit plans and past visit plans to calculate predicted data showing future congestion conditions, and generating a future visit plan for the target user based on the predicted data, it is possible to generate a visit plan to be recommended to the target user, taking into account visit plans recommended to other users.

[0118] Furthermore, by calculating the ratio between past visit plans and the actual number of people, it is possible to calculate forecast data from recommended visit plans in situations where users do not necessarily follow the recommended visit plans.

[0119] Furthermore, the present invention is not limited to the device configuration and operation of the above-described embodiment, and various modifications and applications are possible within the scope of the gist of the present invention.

[0120] For example, the number of users may indicate the number of groups, such as families, rather than the number of people. In this case, the number of people included in the group corresponding to each user may be acquired, and the number of people in each group may be added together to calculate the number of people indicated by the number of users. It may also be calculated as a number of people.

[0121] In addition, although an example has been described in which one-hour units are used as time units, the present invention is not limited to this, and time units may be divided into smaller or larger units than one hour. For example, time units may be monthly, daily, or day-of-the-week units. Note that the more detailed the time units are, the more detailed recommendations can be made, but there is a higher possibility that there will not be enough data. Furthermore, more detailed data may be obtained by combining with other data. For example, by combining with data indicating sunny or rainy days, it is possible to take into account that there are destinations where the number of people visit increases when it rains.

[0122] Furthermore, by bundling multiple destination IDs, the number of people can be calculated by treating them as a single destination. Even if the amount of data is small and trends cannot be identified when only one destination is used, it is expected that trends can be identified by adding up the number of people at multiple destinations. From another perspective, if each destination is visited in a short time, it may not take as long as a period unit, but by treating multiple destinations together, it is expected that the time required to complete a visit can be adjusted to the granularity of a period unit.

[0123] Furthermore, when calculating the ratio, information such as gender and age may be further used to obtain more detailed trends regarding whether or not to follow a visit plan. By enabling more detailed adjustments in this way, it is expected that the error in the number of people predicted can be reduced. While the example described above uses a case in which the ratio between the number of users recommended to visit a destination for each period in past visit plans and the actual number of users who visited the destination for each period is calculated, this is not limiting. It is also possible to calculate predicted data for the number of users using a general machine learning algorithm, without necessarily calculating the ratio, using actual data, past visit plans, future visit plans, gender and age information, and meteorological data such as temperature and weather as feature quantities. However, in this case, it is assumed that the amount of data sufficient for machine learning and correct answer data are available.

[0124] The destination may be a facility, or may be a food stall or food truck that is not a building. Furthermore, the destination of an online event may be used as the destination, and the number of users accessing the event simultaneously may be used.

[0125] In addition, various processes executed by the CPU after reading software (programs) in the above embodiments may be executed by various processors other than the CPU. Examples of such processors include programmable logic devices (PLDs) (such as field-programmable gate arrays (FPGAs)) whose circuit configuration can be changed after manufacture, and dedicated electrical circuits, such as application-specific integrated circuits (ASICs), which are processors with circuit configurations specifically designed to execute specific processes. The plan generation process may be executed by one of these various processors, or by a combination of two or more processors of the same or different types (e.g., multiple FPGAs, or a combination of a CPU and an FPGA). The hardware structure of these various processors is, more specifically, an electrical circuit that combines circuit elements such as semiconductor devices.

[0126] In addition, in each of the above embodiments, the calculation program is pre-stored (installed) in the storage 14, but this is not limiting. The program may be provided in a form stored in a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), or a USB (Universal Serial Bus) memory. The program may also be downloaded from an external device via a network. The technology of the present disclosure may also be applied to a program product.

[0127] The following additional notes are provided regarding the above-described embodiments.

[0128] (Supplementary Item 1) A calculation device including: a memory; and at least one processor connected to the memory, wherein the processor is configured to calculate forecast data representing the number of users expected to visit each of a plurality of destinations for each predetermined time unit, based on performance data representing the actual number of users who have visited each of a plurality of predetermined destinations for each predetermined time unit, and visiting plans generated for each user representing destinations to be recommended to the user for each time unit, the visiting plans corresponding to the performance data, and the visiting plans generated for each user for a future time period beyond the present.

[0129] (Supplementary Item 2) A non-transitory storage medium storing a program executable by a computer to execute a plan generation process, wherein the plan generation process calculates forecast data representing the number of users expected to visit each of a plurality of destinations for each predetermined time unit, based on performance data representing the actual number of users who have visited each of a plurality of predetermined destinations for each predetermined time unit, and visiting plans generated for each user representing destinations recommended to the user for each time unit, the visiting plans corresponding to the performance data, and the visiting plans generated for each user for a future time period beyond the present.

[0130] REFERENCE SIGNS LIST 10 Calculation device 11 CPU 13 RAM 14 Storage 15 Input unit 16 Display unit 20 Performance database 22 Visit plan database 24 Calculation unit 26 Plan generation unit 30 Future plan acquisition unit 32 Number of users calculation unit 34 Past plan acquisition unit 36 ​​Number of users calculation unit 38 Performance data acquisition unit 40 Ratio calculation unit 42 Product calculation unit 44 Proportion calculation unit 46 Sum calculation unit 48 Prediction data calculation unit 50 Setting unit 52 Score calculation unit 54 Selection probability calculation unit 56 Visit plan generation unit 58 Evaluation unit

Claims

1. A calculation device including: performance data representing the actual number of users who have visited each of a predetermined number of destinations for each predetermined time unit; a visiting plan generated for each user representing destinations recommended to the user for each time unit, the visiting plan corresponding to the performance data; and the visiting plan generated for each user representing a future time period from the present time, the calculation unit calculating forecast data representing the number of users expected to visit each of the plurality of destinations for each predetermined time unit.

2. The calculation unit calculates, for each destination and for each period, a ratio between the number of users recommended to visit the destination in the past visit plan for each period and the actual number of users who visited the destination in the period; and calculates, for each destination and for each period, forecast data on the number of users expected to visit the destination in the period based on the calculated ratio between the number of users recommended to visit the destination in the future visit plan for each period and the calculated ratio.

3. The calculation device according to claim 2, wherein the calculation unit calculates, for each destination and for each time period, the product of the number of users recommended to visit the destination in the future visit plan for each time period and the calculated ratio; calculates, for each destination, the ratio of the product for each time period to the sum of the products for each time period; and calculates, for each time period, predicted data on the number of users based on the calculated ratio and the sum of the number of users who visited the destination obtained from the performance data.

4. The calculation device according to claim 1, further comprising a plan generation unit that generates the future visiting plan for the target user based on the forecast data.

5. The calculation device according to claim 4, wherein the plan generation unit generates the future visiting plan based on the forecast data and a recommendation score obtained for each visiting destination.

6. A calculation method implemented by a computer, which calculates forecast data representing the number of users expected to visit each of a plurality of predetermined destinations for each predetermined time unit, based on performance data representing the actual number of users who have visited each of a plurality of predetermined destinations for each predetermined time unit, a visiting plan generated for each user representing destinations recommended to the user for each time unit, the visiting plan corresponding to the performance data, and the visiting plan generated for each user for a future time period from the present.

7. A calculation program that causes a computer to execute the following steps: Calculate forecast data that represents the number of users expected to visit each of a plurality of predetermined destinations for each predetermined time unit, based on performance data that represents the actual number of users who have visited each of a plurality of predetermined destinations for each predetermined time unit, a visiting plan generated for each user that represents destinations recommended to the user for each time unit, and past visiting plans that correspond to the performance data, and the visiting plans generated for each user that are future visiting plans from the present time.

Images