JPH06348724A - Wait time display device - Google Patents

Abstract

PURPOSE:To provide wait time display device which can actualize the estimation of the exclusive use time of a customer by fuzzy inference without requiring a reception clerk who is well experienced and display the wait time up to the acquisition of a resource corresponding to a changing state by estimating the exclusive use time of the customer by using the fuzzy inference. CONSTITUTION:An inference means 113a finds expected use end time from the estimated use time of the customer found by fuzzy inference by using data in a means 11d1 and data in a means 11d4 obtained by a 1st means 14, and stores it in a means 11d2. A display control means 11b finds the wait time of an (N)th resource waiting customer by using data in a means 11d2 and outputs it to a display means 17. A learning means 113c means and standard deviations by the days of the week and time zones by using data in a means 11d3 obtained after operation, and corrects and stores the membership function of the expected use time in the means 11d4.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a waiting time display device, and more particularly, to a customer who has limited resources such as a practice space in a golf driving range, a parking space in a parking lot, a restaurant seat, and a consultation at a hospital. The present invention relates to a device for displaying a waiting time until a customer obtains a resource when it is provided to the customer.

[0002]

2. Description of the Related Art Generally, at golf driving ranges, parking lots, restaurants, etc., the current waiting time is generally notified to customers by the estimated time by the receptionist based on the length of the waiting customers and the number of customers. , An unfamiliar receptionist cannot predict how much time they will have. Also, even in the same situation,
The predicted waiting time will differ depending on the receptionist.

Therefore, the current waiting time is indicated to the customer according to the waiting time of the customer who finally acquired the resource, or, for example, as disclosed in Japanese Patent Laid-Open No. 3-201200, a unit is set in a parking lot. It has been proposed to determine the occurrence status of vacancy from the number of vehicles delivered per hour and calculate the waiting time.

[0004]

However, according to the above-mentioned proposal, in a situation where there is a large change in the usage state of the resource, such as when starting to provide the resource, the resource is acquired without waiting within the capacity of the resource. It is possible, but after the capacity is exceeded, the waiting customer cannot acquire the resource unless the customer has finished using the resource. Under such a situation, the waiting time cannot be obtained unless the customer who has acquired the resource first can know the time when the use is finished. The above-mentioned conventional device has a problem that the waiting time in the above situation cannot be obtained because the customer who has acquired the resource does not predict the time to finish the use.

Therefore, in view of the above-mentioned conventional problems, the present invention embodies the idea by estimating the customer's resource occupation time by using fuzzy inference, and realizes the idea without requiring an experienced receptionist to provide resources. A first object of the present invention is to provide a waiting time display device capable of accurately obtaining and displaying the waiting time until the acquisition of resources even in a situation where the usage state of resources is largely changed such as at the start.

A second object of the present invention is to provide a waiting time display device capable of updating a membership function for fuzzy reasoning about a customer's resource occupation time to an optimum one by learning after operation. I am trying.

[0007]

[Means for Solving the Problems] The above first object is achieved.
Therefore, the waiting time display device made according to the present invention is shown in FIG.
Indicate the customer's resource acquisition time as shown in the basic configuration diagram of
Time data generating means 14 for generating resource acquisition time data
And the resource acquisition time data generated by the time data generating means.
First customer data storage means 11d for storing data₁And
Display means 17 for displaying the waiting time for thinking,
Stores the expected end time of resource use for customers waiting to acquire resources
2 customer data storage means 11d₂And resource acquisition time
Store membership function with expected usage time by time zone
Inference data storage means 11d _FourAnd the first customer
Using the data storage means and the inference data storage means,
Prediction is based on the estimated usage time of the customer, which is obtained by performing a logical reasoning
The second customer data storage for obtaining the end time of using the resource
The inference means 113a stored in the means and the second customer data
The first among customers waiting for resource acquisition using the data storage means
The waiting time of the optional customer from the
And a display control means 113b for outputting.
It is a sign.

The waiting time display device displays the use end detection means 113d for detecting the customer who has finished using the resource and the data in the first customer data storage means corresponding to the customer detected by the use end detection means. It is characterized by further comprising erasing means 113f for erasing.

In order to achieve the second object, the waiting time display device according to the present invention generates the resource acquisition time data indicating the resource acquisition time of the customer as shown in the basic configuration diagram of FIG. No. 1 time data generating means 14, second time data generating means 16 for generating resource use end time data indicating the resource use end time of the customer, display means 17 for displaying expected waiting time, and the time data generation. the first customer data storage means 11d ₁ for storing the resource acquisition time data means generates, as the second customer data storage means 11d ₂ for storing in resource use and expected resource use end time of resource acquisition waiting customers, Based on the time data generated by the first and second time data generating means, a third customer data storing resource acquisition time and resource use end time of a customer who has used the resource in the past. Data storage means 11d ₃ , inference data storage means 11d ₄ for storing membership functions of resource acquisition time zones and expected usage times by time zone, and fuzzy inference using the inference data storage means. Calculate the expected end time of resource use according to the estimated usage time of the customer,
Inference means 11 to be stored in the second customer data storage means
3a and the second customer data storage means are used to obtain the waiting time of an arbitrary customer from the beginning among the customers waiting for resource acquisition and display control means 1 for outputting the waiting time to the display means.
13b, and learning means 113c for obtaining the membership function of the expected usage time by using the third customer data storage means and replacing and updating it with the membership function of the expected usage time in the inference data storage means. Is characterized by.

The learning means may determine the membership function of the expected usage time based on the average and standard deviation of the resource usage time obtained for each time zone from the data obtained in the third customer data storage means after the operation. It has a feature.

[0011]

In the structure which achieves the first object, the time data generating means 14 for generating the resource acquisition time data indicating the resource acquisition time of the customer is provided. The second customer data storage means 11d ₂ contains the expected resource use end times of the customers who are using the resource and who are waiting for the resource acquisition, and the inference data storage means 11d ₄ contains the resource acquisition time zone and the expected use time by time zone. Each ship function is stored.

The inference means 113a obtains the expected resource use end time based on the expected use time of the customer obtained by performing fuzzy inference using the first customer data storage means 11d ₁ and the inference data storage means 11d ₄ Is stored in the second customer data storage means 11d ₂ . Display control means 113
b uses the second customer data storage means 11d ₂ to obtain the waiting time of an arbitrary customer from the beginning among the customers waiting for resource acquisition, and outputs it to the display means 17.

In the configuration achieving the second object, the first time data generating means 14 for generating the resource acquisition time data indicating the resource acquisition time of the customer, and the resource use end time indicating the resource use end time of the customer. A second time data generating means 16 for generating data is provided. In the third customer data storage means 11d ₃ , the resource acquisition time and resource use end time of the customer who used the resource in the past are stored in the second customer data storage means 1d.
1d ₂ is the expected resource use end time of the customer who is using the resource and waiting for the resource acquisition, and the inference data storage means 11d ₄
Stores membership functions of resource acquisition time zone and expected usage time by time zone.

The inference means 113a performs fuzzy inference using the first customer data storage means 11d ₁ and the inference data storage means 11d ₃ to obtain the expected resource use end time based on the customer's expected use time, and obtains this. It is stored in the _second customer data storage means 11d ₂ . Display control means 113b
Uses the _second customer data storage means 11d ₂ to obtain the waiting time of an arbitrary customer from the beginning among the customers waiting for resource acquisition, and outputs it to the display means 17. Learning means 11
3c is a third customer data storage means 1 obtained after the operation.
The membership function of the expected usage time is created using 1d _3, and the one in the inference data storage means 11d ₄ is updated.

Based on the average and standard deviation of the resource usage time obtained by the time zone from the data obtained in the _third customer data storage means 11d ₃ after the creation of the membership function of the expected usage time by the learning means 113c. To do.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 2 shows an embodiment of a waiting time display device according to the present invention applied to a parking lot. In FIG. 2, 11 is a data processor including a personal computer (hereinafter abbreviated as a personal computer), which is a control program thereof. A ROM 111 storing therein, a RAM 112 storing various data in a readable and writable manner, and a main body 11a having an arithmetic unit 113 for performing data processing according to a control program, and a keyboard 1 connected to the main body 11a.
1b, a CRT 11c, and an external memory 11d such as a hard disk.

First time data generation means for generating the warehousing time data in the personal computer main body 11a in response to the customer's warehousing (resource acquisition) detection by the warehousing detection unit 13 via the interface (I / F) circuit 12. Arrival time data generator 14 as an example, and an exit detection unit 15 for detecting the exit of a vehicle
And a delivery time data generator 16 as second time generation means for generating delivery time data in response to the customer's delivery (resource use end) detection by the delivery detection unit 15, and as display means for displaying the waiting time. And the waiting time indicators 17 are connected respectively.

As the warehousing detection unit 13 and the warehousing time data generator 14, as shown in FIG. 3, a parking ticket issuing device 3 for controlling the gate 2 provided at the entrance of the parking lot 1 can be applied. The parking ticket issuing device 3 manages an empty space in the parking lot by managing the number of vehicles in the parking lot, the information about the delivery from the fee settlement / receipt issuing device 4 installed at the exit toll collection point, and the number of parkable vehicles. When there is, the parking ticket can be issued by the customer's operation, and when issuing the ticket, the warehousing time data for printing the warehousing time on the parking ticket is generated, so the warehousing detection is performed by the ticket issuing operation and the parking ticket is printed. Time data can be used as warehousing time data.

Further, as the exit detection unit 15 and the exit time data generator 16, the charge settlement / receipt issuing device 4 provided at the exit of the parking lot 1 can be applied. Since the charge adjustment / receipt issuing device 4 adjusts the parking charge based on the entering time printed on the parking ticket received from the vehicle that is about to leave and the current time, the time at the time of this adjustment can be used as the exit time data. it can.

In the external memory 11d, a first customer database serving as a first customer data storage means for storing the warehousing time of the warehousing customer based on the warehousing time data generated by the warehousing time data generator 14. 11d ₁ and a second customer database 11 as a second customer data storage means for storing the expected leaving time (resource use end time) of customers who are parking (using resources) and waiting for parking (waiting for resources).
A third customer that stores the warehousing time and warehousing time of a customer who has been used in the past, based on d ₂ and the warehousing time data generated by the warehousing time data generator 14 and the warehousing time data generated by the warehousing time data generator 16. A third customer database 11d ₃ as a data storage means, and an inference database 11d ₄ as an inference data storage means for storing a membership function of an expected parking time (usage time) for each warehousing time zone, day of the week, and time zone. Is provided.

The first customer database 11d ₁ is
As shown in FIG. 4, there is a capacity for the number of cars that can be parked, and the warehousing time data for the vehicles that have left the garage is deleted for each garage leaving.
Instead, the storage time data of the newly stored vehicle is stored. The second customer database 11d ₂ is shown in FIG.
As shown in (a), an area A that stores expected shipping time data for the number of parkable cars + an arbitrary number of cars N, and waiting time data for the number of parkable cars + an arbitrary number of cars N as shown in FIG. 5B. And an area B for storing As shown in FIG. 6, the third customer database 11d ₃ stores, as a pair, data indicating the entering and exiting times of M for an arbitrary number of vehicles actually parked.

The waiting time indicator 17 is, as shown in FIG. 3, a plurality of indicators installed at positions corresponding to a predetermined number of vehicles waiting to park along a line starting from the entrance of the parking lot 1. Each of the indicators is configured to include at least the time the vehicles lined up after the indicator have to wait.

The arithmetic unit 113 operates the first customer database 11 according to the program stored in the ROM 111.
second inference unit 113a as inference means for inferring the expected departure time from the arrival time data stored in d ₁ and the membership function in the inference database 11d ₄ .
Display control section 113b as display control means for obtaining the expected waiting time from the expected parking time in the customer database 11d _{2 of} No. 3, and displaying it on the waiting time indicator 17 in the third customer database 11d ₃ obtained after the inference rule is operated. Of the customer, learns based on the data of each customer to create a membership function, and replaces the existing one with the previous one to update as a learning unit 113c.

The inference unit 113a performs fuzzy inference using the data stored in the first customer database 11d ₁ and the inference database 11d ₄ to obtain the expected parking time and the expected delivery time of the customer, Stored in the customer database 11d ₂ . The display control unit 113b obtains the waiting time of the Nth customer from the head among the customers waiting for parking using the second customer database 11d ₂ and outputs the waiting time to the waiting time indicator 17. Learning unit 113c is, using a third of the customer database 11d ₃ obtained after the operation, day of the week, in search of the mean and standard deviation of parking time in each time zone, to create a prediction parking time membership function, which earlier It is stored in the inference database 11d ₄ so as to be replaced with and updated.

The inference database 11d ₄ stores a membership function for the warehousing time zone and a membership function for the expected parking time. As shown in FIG. 7, the membership function of the warehousing time zone is shown in FIG. 7. The horizontal axis represents time, and the time zone is from 5:00 to 9:00 in the early morning hours, from 8:00 to 13:00 in the morning hours, and at 12:00. -Fitness is trapezoidal for each of the afternoon hours from 17:00 to 16:00, the evening hours from 16:00 to 20:00, the evening hours from 19:00 to 23:00, and the late hours from 22:00 to 6:00 It is set. The shape of the membership function may be, for example, a rod shape or a triangular shape other than the trapezoidal shape.

On the other hand, the membership function of the expected parking time is set for each day of the week and time zone as shown in FIG. 8, and specifically, for the X vehicles obtained during the operation stored in the _third customer database 11d _3. When the parking time data of S is used as an element, the average Tstay and the standard deviation Sstay of the parking time are used, and as shown in FIG.
A trapezoidal shape is created with the point where ± stay is ± as the upper side and the point where ± 3S stay is ± as the lower side. Of course, this membership function may be created by another method, and may be, for example, a rod shape or a triangular shape other than the trapezoid.

Next, the inference unit 113a and the display control unit 113
b and the processing performed by the arithmetic unit 113 constituting the learning unit 113c according to the control program will be described below with reference to the flowcharts of FIGS. First, the process of the flowchart of FIG. 10 is started by turning on the power of the main body 11, and in the first step S1, a timer of, for example, 5 minutes formed in the RAM 112 is started, and then the process proceeds to step S2, where there is a receipt. It is determined whether or not the warehousing time data is input from the warehousing time data generator 14. When the determination in step S2 is NO, the process proceeds to step S3, and it is determined whether there is a delivery or not depending on whether or not there is a signal from the delivery detection unit 15. When the determination in step S3 is NO, the process proceeds to step S4, and it is determined whether or not the time-over flag of the predetermined area in the RAM 112 is turned on.
When the determination in 4 is NO, the process proceeds to step S5, and it is determined whether or not the timer started in step S1 times out. When the determination in step S5 is also NO, the process returns to step S2 and the above steps are repeated.

When the determination in step S2 is YES, that is, when it is determined that the data has been input from the storage time data generator 14 and the storage has occurred, the process proceeds to step S6, in which the storage time data is set to the first customer. After storing in the database 11d, the process proceeds to step S3. When the determination in step S3 is YES, that is, when there is a warehousing, the process proceeds to step S7 to store the warehousing time data in the third customer database 11d ₃ and at the same time warehousing the vehicle in the _first customer database 11d _1. After deleting the time data, the process proceeds to step S8 described later. When the determination in step S4 is YES, that is, when the time-over flag is on, the process proceeds to step S9 to turn off the time-over flag and then to step S8.

In step S8, the expected shipping time data of FIG. 5A and the waiting time data of FIG. 5B stored in the _second customer database 11d ₂ in the external memory 11d are both cleared. After initialization, the process proceeds to step S10. In step S10, the expected exit time of each vehicle is calculated by fuzzy inference from the entry time of the vehicle currently parked based on the entry time data in the _first customer database 11d ₁ , and this is calculated in the second customer database 11d. _It is stored in the area A in ₂ (FIG. 5A).
Then, it progresses to step S11 and it is determined whether there is an empty space in the parking lot. This determination is performed, for example, based on whether or not the difference between the number of stored vehicles and the number of delivered vehicles is equal to or greater than the number of parkable vehicles.

When the determination in step S11 is YES, that is, when there is a vacant space, the process proceeds to step S12, in which the expected time of vacant parking is estimated by fuzzy reasoning assuming that the current time is the warehousing time. The calculated value is stored in the area A in the _second customer database 11d _{2 and then the} process proceeds to step S13. At this time, the waiting time data is stored in area B (FIG. 5B), but the number of vacant parking vehicles is sequentially stored from the beginning as “waiting time = 0.” If the determination in step S11 is NO, skip step S12. And proceeds to step S13.

In step S13, the earliest prediction
The expected shipping time is fuzzy, assuming that the shipping time is the incoming time.
Calculated by inference, this is the second customer database 1
1d ₂Stored additionally. At this time, the assumed storage time and
Waiting time is the difference from the current time
It is additionally stored in area B. Similarly, this process is expected
By repeating the earliest time N times in sequence, the current
To (N + number of available parking spaces) waiting for customers
You can ask for the time. Then proceed to step S14
Here, the second customer database 11d₂Waiting inside
Waiting time for the Na and Nnth items you want to display according to the inter-data
Is displayed on the waiting time indicators 17a and 17n, and then
Return to step S1.

When the determination in step S5 is YES, that is, when the timer has timed out after 5 minutes from its activation, the process goes to step S15 to turn on the time-over flag and then returns to step S1. When the time-over flag is turned on in step S15, the determination in step S4 is YES and the process proceeds to step S8 via step S9.

In order to infer the expected leaving time, the expected parking time Fstay is inferred from the day of the week and the entering time. According to this inference rule, first, a warehousing time zone that matches the warehousing time and its matching degree are calculated. Next, the expected parking time corresponding to the day of the week of the day and the obtained warehousing time zone is retrieved from the database.
Then, the predicted parking time is obtained as a conclusion from the extracted predicted parking time and the previously determined fitness by using a fuzzy calculation (for example, the center of gravity method).

Finally, the process of the flow chart of FIG. 11 is automatically started at the operation of a specific key of the keyboard 11b or at a certain time, and the calculation unit 113, in step S21 immediately after the start of the operation, calculates the third customer database. 11d
Starting from ₃ , the parking time for X cars is read for each day of the week and warehousing time, then the process proceeds to step S22, where the average and standard deviation of the parking time are calculated for each day of the week and warehousing time to create a membership function for the estimated parking time. Then, this is replaced with a previous one and stored in the inference database 11d ₄ for correction, and the learning process ends.

Explaining further with a concrete example, now,
When customer data of the arrival time of 8:30 on Monday is obtained for the car A by the arrival time data generator 13, the inference database is used to obtain the arrival time zone that matches the arrival time as shown by the dotted line in FIG. . In this example, the corresponding warehousing time zones have both goodness of fit for early morning and morning.
It becomes 5. Then, the corresponding expected parking time is retrieved from the inference database. In this example, FIGS. 12A and 12B are used.
As shown in, the membership function of the predicted parking time is 0.5, which has an average fitness of 0.5 hours and 4 hours on Monday, and early morning and morning fitness of 0.5. After that, the expected parking time is calculated by fuzzy calculation. In this example, when the membership functions of FIGS. 12A and 12B are overlapped as shown in FIG.
It will be 4.5 hours.

In the above-mentioned embodiment of displaying the waiting time of the parking lot, paying attention to the fact that the number of vehicles waiting and the length of the queue are substantially proportional, and as shown in FIG.
The waiting time indicators 17a to 17n are installed at a plurality of positions such as positions corresponding to the 0th car and the approximately 20th car. However, when the waiting time display device of the present invention is applied to a golf driving range or a restaurant, Requires some changes.

At the golf practice range, fuzzy inference is performed using the day of the week and the golfer's entry time to estimate the expected departure time of the golfer currently in practice, the availability of practice space and the expected departure time of each golfer. The expected waiting time is displayed based on. The inference rules used for fuzzy inference are redefined by learning the data of each golfer obtained during the operation of the device. The entrance time data should be generated when the practice space used by the front desk is specified, and the exit time data should be generated when the fare is settled, and the display is the number of people currently waiting for practice + 1st waiting time. Should be displayed in.

In the restaurant, fuzzy inference is performed using the day of the week of the day and the customer's entrance time to estimate the expected opening time of the customer who is currently having a meal, and also the availability of seats and the estimated opening time of each customer. And the expected waiting time is displayed. The inference rules used for fuzzy inference are redefined by learning the data of each customer obtained during the operation of the device. The entrance time data may be generated when the waitress receives an order, and the exit time data may be generated when the charge is settled.

Further, if there is a customer's reception card, etc., if the customer recognizes the number of the queue from the card or inputs the number of the queue from the keyboard or the like, the arbitrary number You can also display the waiting time of.

In the embodiment, the inference is made from the day of the week and the resource acquisition time, but by using other elements such as weather and holidays, it is possible to infer the resource use time closer to the actual time.

Further, the above-mentioned device can be economically operated in the network by forming the network.
For example, by networking a plurality of sales offices in the area and displaying the waiting time of each, it becomes possible to operate the entire area efficiently.

[0042]

As described above, according to the present invention, by estimating the customer's resource occupation time using fuzzy inference, the idea can be embodied without requiring an experienced receptionist. It is possible to obtain the effect that the waiting time until the resource acquisition corresponding to the changing situation can be displayed more accurately.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic configuration of a waiting time display device according to the present invention.

FIG. 2 is a diagram showing an embodiment of the device of the present invention applied to a parking waiting time display device.

FIG. 3 is a diagram showing a planar configuration example of a parking lot.

FIG. 4 is a diagram showing a configuration example of a first customer database.

FIG. 5 is a diagram showing a configuration example of a second customer database.

FIG. 6 is a diagram showing a configuration example of a third customer database.

FIG. 7 is a diagram showing a membership function in a time zone.

FIG. 8 is a diagram showing a membership function of parking time by day of the week and time zone.

FIG. 9 is a diagram for explaining how to obtain the function of FIG.

FIG. 10 is a flowchart showing a part of processing performed by a calculation unit in FIG.

FIG. 11 is a flowchart showing a learning process performed by a calculation unit in FIG.

FIG. 12 is a diagram showing a function extracted for predicting a parking time in the inference unit.

FIG. 13 is a diagram showing how to obtain an estimated parking time by fuzzy calculation.

[Explanation of symbols]

11d ₁ 1st customer data storage means (1st customer database) 11d ₂ 2nd customer data storage means (2nd customer database) 11d ₃ 3rd customer data storage means (3rd customer database) 11d ₄ Inference data storage means (inference database) 113a Inference means (inference section) 113b Display control means (display control section) 113c Learning means (learning section) 113d End-of-use detecting means (calculation section) 113f Erasing means (calculation section) 14 First time data generating means (arrival time data generator) 16 Second time data generating means (delivery time data generator) 17 Display means (waiting time indicator)

Claims (4)

[Claims] 1. A time data generation means for generating resource acquisition time data indicating a resource acquisition time of a customer, a first customer data storage means for storing resource acquisition time data generated by the time data generation means, and a prediction Members for displaying a waiting time, a second customer data storing means for storing expected resource use end times of customers who are using resources and waiting for resource acquisition, and a resource acquisition time zone and an expected usage time by time zone An inference data storage means for storing a ship function, and an expected resource use end time is obtained from the expected usage time of the customer obtained by performing fuzzy inference using the first customer data storage means and the inference data storage means. An inference means to be stored in the second customer data storage means, and an arbitrary number customer from the beginning among the customers waiting for resource acquisition using the second customer data storage means For the waiting time of
A waiting time display device, comprising: a display control means for outputting to the display means. 2. A use end detecting means for detecting a customer who has finished using the resource, and an erasing means for erasing data in the first customer data storing means corresponding to the customer detected by the use end detecting means. The waiting time display device according to claim 1, further comprising: 3. A first time data generating means for generating resource acquisition time data showing a customer's resource acquisition time, and a second time data generating means for generating resource use end time data showing a customer's resource use end time. Means, a display means for displaying an expected waiting time, a first customer data storage means for storing the resource acquisition time data generated by the time data generating means, a resource end in use and an expected resource end of a customer waiting for resource acquisition Second customer data storage means for storing the time, and based on the time data generated by the first and second time data generating means, the resource acquisition time and the resource use end time of the customer who has used the resource in the past. Third customer data storage means for storing, inference data storage means for storing a resource acquisition time zone and a membership function of expected use time by time zone, and the inference data storage means Data storage means, fuzzy inference is performed to obtain the expected use time of the customer based on the expected use time of the customer, and the inference means stores the estimated resource use end time in the second customer data storage means, and the second customer data storage means. Using, to find the waiting time of any customer from the beginning among the customers waiting for resource acquisition,
Using the display control means for outputting to the display means and the third customer data storage means, a membership function of the expected usage time is obtained, and the membership function of the expected usage time in the inference database is replaced and updated. A waiting time display device comprising: 4. The learning means determines the membership function of the expected usage time based on the average and standard deviation of the resource usage times obtained for each time zone from the data obtained in the third customer data storage means after the operation. 4. The waiting time display device according to claim 3, wherein: