Patent Application
20240012682
The disclosed technology relates to an allocation search device, allocation search method, and allocation search program.
There is known a technology of predicting occurrence of an event in each region and optimally allocating a plurality of resources for the event. For example, Non Patent Literature 1 proposes a method of predicting a regional emergency demand and optimally allocating a plurality of ambulance squads (ambulances) on the basis of the prediction so that a time required to arrive at a site or a travel distance required to arrive at a site is reduced as much as possible.
For example, there is a case where occurrence of a sick/injured person per unit time is predicted in each region mesh of 500 m square or 1 km square, and a plurality of ambulance squads (ambulances) is appropriately allocated so that a time required to arrive at a site or a travel distance required to arrive at a site is reduced as much as possible. The ambulance squads can be allocated in a plurality of fire stations existing in a target region, and one ambulance squad waiting at the shortest distance from a site of occurrence of a sick/injured person is dispatched. In this case, the plurality of ambulance squads is to be appropriately allocated in the plurality of fire stations. Thus, a discrete optimization problem arises.
However, for example, there are about 50 fire stations and about 40 ambulance squads in core cities in Japan. In this case, for example, when it is simplified that any number of ambulance squads can be allocated in each fire station, the number of allocation patterns is 50 to the 40th power by simple calculation. Therefore, it is difficult to find an optimal solution of such a problem in real time.
Meanwhile, even if occurrence of a sick/injured person is predicted in each regional mesh, a result thereof may be wrong. Regarding this point, it is desirable that ambulance squads be in robust allocation that can expect an effect even if the prediction is not completely right.
The disclosed technology has been made in view of the above points, and an object thereof is to provide an allocation search device, allocation search method, and allocation search program capable of obtaining, in a short time during which it is necessary to determine effective allocation of resources, robust allocation of the resources that can expect an effect even if a prediction result of occurrence of an event is not completely right.
In order to achieve the above object, an allocation search device according to an aspect of the present disclosure includes: a first sample output unit configured to output, on the basis of event occurrence data obtained in the past, event occurrence data that may occur under a predetermined condition as a first main sample and event occurrence data that may occur under a condition similar to the predetermined condition as a first auxiliary sample; an allocation plan creation unit configured to create a plurality of allocation plans of resources for an occurring event; a first allocation evaluation unit configured to evaluate whether or not each of the plurality of allocation plans satisfies a predetermined evaluation criterion in a case where the first main sample and the first auxiliary sample are applied to each of the plurality of allocation plans; a second sample output unit configured to output, on the basis of latest event occurrence data, future event occurrence data that may occur under the predetermined condition as a second main sample and future event occurrence data that may occur under the similar condition as a second auxiliary sample; and a second allocation evaluation unit configured to, in a case where the second main sample and the second auxiliary sample are applied to each of the allocation plans that satisfy the predetermined evaluation criterion, reevaluate whether or not each of the allocation plans that satisfy the predetermined evaluation criterion satisfies the predetermined evaluation criterion.
In order to achieve the above object, an allocation search method according to an aspect of the present disclosure includes: outputting, on the basis of event occurrence data obtained in the past, event occurrence data that may occur under a predetermined condition as a first main sample and event occurrence data that may occur under a condition similar to the predetermined condition as a first auxiliary sample; creating a plurality of allocation plans of resources for an occurring event; evaluating whether or not each of the plurality of allocation plans satisfies a predetermined evaluation criterion in a case where the first main sample and the first auxiliary sample are applied to each of the plurality of allocation plans; outputting, on the basis of latest event occurrence data, future event occurrence data that may occur under the predetermined condition as a second main sample and future event occurrence data that may occur under the similar condition as a second auxiliary sample; and, in a case where the second main sample and the second auxiliary sample are applied to each of the allocation plans that satisfy the predetermined evaluation criterion, reevaluating whether or not each of the allocation plans that satisfy the predetermined evaluation criterion satisfies the predetermined evaluation criterion.
In order to achieve the above object, an allocation search program according to an aspect of the present disclosure causes a computer to execute: outputting, on the basis of event occurrence data obtained in the past, event occurrence data that may occur under a predetermined condition as a first main sample and event occurrence data that may occur under a condition similar to the predetermined condition as a first auxiliary sample; creating a plurality of allocation plans of resources for an occurring event; evaluating whether or not each of the plurality of allocation plans satisfies a predetermined evaluation criterion in a case where the first main sample and the first auxiliary sample are applied to each of the plurality of allocation plans; outputting, on the basis of latest event occurrence data, future event occurrence data that may occur under the predetermined condition as a second main sample and future event occurrence data that may occur under the similar condition as a second auxiliary sample; and, in a case where the second main sample and the second auxiliary sample are applied to each of the allocation plans that satisfy the predetermined evaluation criterion, reevaluating whether or not each of the allocation plans that satisfy the predetermined evaluation criterion satisfies the predetermined evaluation criterion.
According to the disclosed technology, it is possible to obtain, in a short time during which it is necessary to determine effective allocation of resources, robust allocation of the resources that can expect an effect even if a prediction result of occurrence of an event is not completely right.
Hereinafter, an example of an embodiment of the disclosed technology will be described with reference to the drawings. In the drawings, the same or equivalent components and portions will be denoted by the same reference signs. Further, dimensional ratios in the drawings are exaggerated for convenience of description and thus may be different from actual ratios.
In this embodiment, there will be described an aspect in which a regional emergency demand is predicted and a plurality of ambulance squads (ambulances) is optimally allocated on the basis of the prediction so that a time required to arrive at a site or a travel distance required to arrive at a site is reduced as much as possible. However, this embodiment can be applied as long as resources can be optimally allocated for an occurring event.
As illustrated in
The CPU 11 is a central processing unit, and executes various programs and controls each unit. That is, the CPU 11 reads the programs from the ROM 12 or the storage 14 and executes the programs by using the RAM 13 as a work area. The CPU 11 controls each component described above and performs various types of operation processing according to the programs stored in the ROM 12 or the storage 14. In this embodiment, the ROM 12 or the storage 14 stores an allocation search program for searching for optimal allocation of resources.
The ROM 12 stores various programs and various types of data. The RAM 13 temporarily stores the programs or data as a work area. The storage 14 includes a hard disk drive (HDD) or a solid state drive (SSD) and stores various programs including an operating system and various types of data.
The input unit 15 includes a pointing device such as a mouse and a keyboard and is used to perform various inputs to the allocation search device.
The display unit 16 is, for example, a liquid crystal display and displays various types of information. The display unit 16 may function as the input unit 15 by adopting a touchscreen system.
The communication interface 17 is an interface through which the allocation search device communicates with another external device. The communication is performed in conformity to, for example, a wired communication standard such as Ethernet (registered trademark) or fiber distributed data interface (FDDI) or a wireless communication standard such as 4G, 5G, or Wi-Fi (registered trademark).
For example, a general-purpose computer device such as a server computer or personal computer (PC) is applied to the allocation search device 10 according to this embodiment.
Next, functional configurations of the allocation search device 10 will be described with reference to FIG. 2.
As illustrated in
The first sample output unit 101 includes a first main sample output unit 101A and a first auxiliary sample output unit 101B, and the second sample output unit 104 includes a second main sample output unit 104A and a second auxiliary sample output unit 104B.
The storage 14 stores ambulance data 141, fire station data 142, past event occurrence data 143, an effective allocation plan 144, and latest event occurrence data 145. The ambulance data 141, the fire station data 142, the past event occurrence data 143, the effective allocation plan 144, and the latest event occurrence data 145 may be stored in an external storage device.
The past event occurrence data 143 is a data row of event occurrence data obtained in the past. The past herein means a certain period in the past from a current point of time at which allocation search is performed and is, for example, a period of past several months or past several years. The latest event occurrence data 145 is a data row of the latest event occurrence data. The latest herein means a certain period immediately before a current point of time at which allocation search is performed and is, for example, a period of the last several days or the last several months. That is, the latest period is shorter than the past period. The event occurrence data is, for example, sick/injured person occurrence data (i.e., data indicating the year, month, date, day, hour, minute, longitude, and latitude at which a sick/injured person occurs).
Based on the past event occurrence data 143, the first sample output unit 101 outputs event occurrence data that may occur under a predetermined condition as a first main sample and outputs event occurrence data that may occur under a condition similar to the predetermined condition as a first auxiliary sample. The predetermined condition is, for example, a condition such as 10:00 on weekdays in September, and the condition similar to the condition is, for example, a condition such as 10:00 on weekdays in August and October. In this embodiment, the first main sample output unit 101A outputs the first main sample, and the first auxiliary sample output unit 101B outputs the first auxiliary sample.
The first main sample may be expressed as, for example, a pseudo occurrence data row generated in a pseudo manner in accordance with an event occurrence frequency that is obtained in each certain area on the basis of a data row that actually occurs under the predetermined condition. Similarly, the first auxiliary sample may be expressed as, for example, a pseudo occurrence data row generated in a pseudo manner in accordance with an event occurrence frequency that is obtained in each certain area on the basis of a data row that actually occurs under the condition similar to the predetermined condition. The first main sample may also be expressed as a pseudo occurrence data row generated in a pseudo manner in accordance with a random variable of an event occurrence probability that is obtained on the basis of the data row that actually occurs under the predetermined condition. Similarly, the first auxiliary sample may also be expressed as a pseudo occurrence data row generated in a pseudo manner in accordance with a random variable of an event occurrence probability that is obtained on the basis of the data row that actually occurs under the condition similar to the predetermined condition. Specific examples of the event occurrence frequency and the event occurrence probability will be described later.
The allocation plan creation unit 102 creates a plurality of allocation plans of resources for an occurring event. In this embodiment, the allocation plan creation unit creates a plurality of allocation plans of ambulances for allocating the ambulances to fire stations by using the ambulance data 141 and the fire station data 142.
The first allocation evaluation unit 103 uses the first main sample, the first auxiliary sample, and the plurality of allocation plans as inputs and evaluates whether or not each of the plurality of allocation plans satisfies a predetermined evaluation criterion in a case where the first main sample and the first auxiliary sample are applied to each of the plurality of allocation plans. The first allocation evaluation unit 103 stores an allocation plan that satisfies the predetermined evaluation criterion in the storage 14 as the effective allocation plan 144. The above processing is “phase 1”.
Next, based on the latest event occurrence data 145, the second sample output unit 104 outputs future event occurrence data that may occur under a predetermined condition as a second main sample and outputs future event occurrence data that may occur under a condition similar to the predetermined condition as a second auxiliary sample. Both the predetermined condition and the condition similar to the predetermined condition are the same as the conditions in the first sample output unit 101. In this embodiment, the second main sample output unit 104A outputs the second main sample, and the second auxiliary sample output unit 104B outputs the second auxiliary sample.
The second allocation evaluation unit 105 uses the second main sample, the second auxiliary sample, and the effective allocation plan 144 as inputs and reevaluates whether or not each effective allocation plan 144 satisfies a predetermined evaluation criterion in a case where the second main sample and the second auxiliary sample are applied to each effective allocation plan 144.
As a result of the reevaluation by the second allocation evaluation unit 105, the result output unit 106 outputs an effective allocation plan that satisfies the predetermined evaluation criterion as optimal allocation of the resources. The above processing is “phase 2”.
This embodiment roughly includes two implementation phases, i.e., phase 1 and phase 2, as described above. Phase 1 is a phase in which a large number of allocation patterns in effective ambulance allocation plans are found in advance on the basis of past sick/injured person occurrence data. Phase 1 is performed, for example, at the beginning of the year, every quarter, or once a month. Phase 2 is a phase in which, for example, occurrence of a sick/injured person in near future is predicted on the basis of the latest sick/injured person occurrence data at the same time every day, the most effective allocation pattern is found from the effective allocation patterns found in advance in the phase 1, and allocation is changed according to the most effective allocation pattern.
First, the processing in phase 1 will be described by exemplifying a case where, for example, optimal allocation of ambulances between 10:00 and 11:00 on weekdays in the next month September is obtained in August. Based on the past sick/injured person occurrence data accumulated by August, the first main sample output unit 101A outputs a plurality of occurrence data rows predicted to be most likely to occur under a predetermined condition (e.g. between 10:00 and 11:00 on weekdays in September) as the first main sample. Further, based on the past sick/injured person occurrence data accumulated by August, the first auxiliary sample output unit 101B outputs a plurality of occurrence data rows predicted to be likely to occur under a condition similar to the predetermined condition (e.g. between 10:00 and 11:00 on weekdays in August and October) as the first auxiliary sample.
There is a plurality of methods of outputting the first main sample and the first auxiliary sample. A first method, which is the simplest method, is to output actual occurrence data rows between 10:00 and 11:00 on weekdays in September for the past several years as they are as the first main sample and output actual occurrence data rows between 10:00 and 11:00 on weekdays in August and October for the past several years as the first auxiliary sample. The samples are processed as described above on the following two assumptions: similar sick/injured person occurrence patterns occur in the same month, the same day, and the same time slot every year; and the occurrence data rows in August and October are similar to the occurrence data rows in September because, for example, average daily temperatures in August and October are relatively close to that in September.
The sick/injured person occurrence data rows in
As a second method, which is another simple method, actual occurrence data rows between 10:00 and 11:00 on weekdays in September for the past several years may be output as they are as the first main sample, and actual occurrence data rows between 10:00 and 11:00 on weekdays from January to December including a period of time between 10:00 and 11:00 on weekdays in September for the past several years may be output as the first auxiliary sample.
As a third method, which is still another method, a method of creating and using pseudo occurrence data rows will be described. Specifically, the first main sample output unit 101A obtains a sick/injured person occurrence frequency in a certain area (e.g. every 500 m square or every 1 km square) on the basis of the actual occurrence data rows between 10:00 and 11:00 on weekdays in September for the past several years and generates pseudo occurrence data rows in accordance with the frequency. The sick/injured person occurrence frequency is an example of the event occurrence frequency. For example, the sick/injured person occurrence frequency in a certain area of 500 m square between 10:00 and 11:00 on weekdays in September is obtained as 30/100=0.3, where the past occurrence data rows to be used are, for example, data for 100 days and the total number of occurrence of sick/injured people during the period is, for example, 30 people. Similarly, the occurrence frequency may be obtained in all areas of a target region, pseudo occurrence data rows corresponding to values of the occurrence frequencies may be generated, and the data rows in all the areas may be used as one set. Regarding in which position in the target area a sick/injured person occurs, a sick/injured person may occur on the basis of, for example, a density obtained by kernel density estimation that is performed by plotting past actual occurrence positions. An advantage of the method of creating pseudo occurrence data rows is that robust verification using more occurrence patterns can be performed by creating more samples than occurrence data rows that have actually occurred in the past.
At this time, the auxiliary sample output unit 101B may increase or decrease 0.3 that is a value of the occurrence frequency calculated as described above by a certain value (e.g. increase or decrease 0.3 by 0.05 every time, thereby obtaining values of 0.35 and 0.25) and generate pseudo occurrence data rows on the basis of the values. Alternatively, as in the above example, the auxiliary sample output unit may obtain the occurrence frequency on the basis of the past occurrence data rows in August and October and then generate pseudo occurrence data rows.
The pseudo occurrence data rows in
The pseudo occurrence data rows in
As a fourth method, which is further another method, a method of obtaining a sick/injured person occurrence probability as a random variable and creating pseudo occurrence data will be described. The sick/injured person occurrence probability is an example of the event occurrence probability. Obtaining the sick/injured person occurrence probability as the random variable means that a possibility that the sick/injured person occurrence probability takes various values is considered, for example, a possibility that 0.3 people occur is 50%, a possibility that 0.31 people occur is 10%, a possibility that 0.32 people occur is 5%, . . . , and each possibility is expressed as a probability. In order to obtain the sick/injured person occurrence probability as the random variable, target past occurrence data rows may be assumed to occur according to the Poisson distribution, and a parameter of the Poisson distribution (indicating how many times a sick/injured person occurs within a certain period of time) may be obtained as the random variable by using the Markov chain Monte Carlo method (MCMC) or the like. From this result, the first main sample output unit 101A may generate pseudo occurrence data on the basis of a parameter having the highest probability, and the first auxiliary sample output unit 101B may generate pseudo occurrence data on the basis of parameters having other probabilities. In practice, the occurrence probability has a continuous distribution, and thus the pseudo occurrence data may be generated by picking up parameters at certain intervals from the distribution.
The graphs in
In a case where both the first main sample output unit 101A and the first auxiliary sample output unit 101B generate pseudo occurrence data rows, the number of days for the first auxiliary sample output unit 101B may be intentionally reduced. However, the length of the data rows is not reduced. The length of the data rows increases as the parameter of the occurrence probability increases. In a case where the parameter of the Poisson distribution is obtained as the random variable by MCMC described above, the number of days of data generation may be determined according to the magnitude of the probability. In the example of
The occurrence data rows created by the first main sample output unit 101A and the first auxiliary sample output unit 101B are output to the first allocation evaluation unit 103.
Meanwhile, the allocation plan creation unit 102 creates, for example, a plurality of allocation plans for allocating ambulances to fire stations on the basis of the ambulance data 141 in
In the example of the ambulance data 141 in
The allocation plans created by the allocation plan creation unit 102 are output to the first allocation evaluation unit 103. There are various methods of creating the second and subsequent allocation plans. The simplest method is a method of also randomly creating the second and subsequent allocation plans. However, in this case, in a case where the number of combinations is enormous, a long time may be required until an allocation plan evaluated to be effective by the first allocation evaluation unit 103 is specified. In view of this, various heuristics (also referred to as heuristic methods) can be used. One of the methods is a method using a genetic algorithm.
An example of the method using the genetic algorithm will be described. An allocation plan is randomly created until an allocation plan evaluated to be effective by the first allocation evaluation unit 103 is specified, and, in a case where an allocation plan is evaluated to be effective, a next allocation plan is created by randomly changing a part of the allocation plan on the basis of the allocation plan or combining a plurality of allocation plans evaluated to be effective. Combining the plurality of allocation plans means that, for example, allocation of the ambulances a to c is extracted from one allocation plan, allocation of the ambulances d to f is extracted from another allocation plan, and the allocations are combined. In this way, it is empirically known that a solution close to an optimal solution can be obtained in a relatively short time.
Next, the first allocation evaluation unit 103 evaluates the allocation plans acquired from the allocation plan creation unit 102 by using the occurrence data rows acquired from both the first main sample output unit 101A and the first auxiliary sample output unit 101B. As an evaluation method, for example, an average of distances that dispatchable ambulances existing closest to a site of occurrence of a sick/injured person travel until the ambulances arrive at the site is calculated, and the calculated average travel distance is compared with, for example, an average travel distance in original ambulance allocation in
In a case where a certain ambulance is dispatched, the certain ambulance cannot respond to the next request for dispatch for a certain period of time. The certain period of time may be, for example, a value of an average time required to complete transport of a sick/injured person and obtained in advance. For example, in a case where the average time required to complete transport of a sick/injured person is 50 minutes, the certain ambulance cannot respond to the next request for dispatch for 50 minutes. As a method of obtaining a distance from a site of occurrence of a sick/injured person, for example, a direct distance may be used most simply, or the shortest distance on a road network may be used in a case where road network data can be prepared.
In a case where the first main sample output unit 101A outputs data for 100 days, for example, the first allocation evaluation unit 103 evaluates all the data for 100 days. The same applies to the data output from the first auxiliary sample output unit 101B.
As initial values of states of the ambulances, in practice, there is a possibility that several ambulances are already dispatched. Therefore, for example, it is desirable to perform evaluation in various initial states for each sample for one day. The initial states are, for example, a state in which the ambulance a is currently dispatched and returns in 30 minutes and a state in which the ambulances a and b are currently dispatched, the ambulance a can respond in 30 minutes, and the ambulance b can respond in 40 minutes.
As a result of the evaluation described above, in a case where an allocation plan satisfies the predetermined evaluation criterion, the allocation plan is regarded as effective and is stored in the storage 14 as the effective allocation plan 144.
At this time, basically, the occurrence data rows acquired from the first main sample output unit 101A and the occurrence data rows acquired from the first auxiliary sample output unit 101B have different importance and therefore may be evaluated on the basis of different evaluation criteria. For example, in the occurrence data rows acquired from the first main sample output unit 101A, the allocation plan satisfies the evaluation criterion in a case where an average distance required to arrive at a site (hereinafter, referred to as a “site arrival distance”) is shortened from an average distance in original allocation serving as a reference by 100 m or more on average. Meanwhile, in the occurrence data rows acquired from the first auxiliary sample output unit 101B, the allocation plan satisfies the evaluation criterion in a case where the average site arrival distance is shortened by 50 m or more.
In the effective allocation plans 144 in
Meanwhile, the allocation plans in
In a case where the number of days in the occurrence data rows output by the first auxiliary sample output unit 101B is intentionally reduced to be smaller than the number of days output by the first main sample output unit 101A as described above, evaluation may be performed with a weighted average on the basis of a unified evaluation criterion.
The first allocation evaluation unit 103 may output an evaluation result of a certain allocation plan to the allocation plan creation unit 102 and reflect the evaluation result in creation of the next allocation plan.
As described above, it is possible to empirically find several tens to several hundreds of effective allocation patterns by repeating the above processing in phase 1 for more than ten hours by using a general-purpose PC or the like.
Next, an operation of the allocation search device 10 according to this embodiment will be described with reference to
In step S101 of
In step S102, the CPU 11 outputs event occurrence data that may occur under a predetermined condition as a first main sample on the basis of the past event occurrence data 143 accepted as the input in step S101. In this example, the predetermined condition indicates, for example, a condition such as a period of time between 10:00 and 11:00 on weekdays in September as described above.
In step S103, the CPU 11 outputs event occurrence data that may occur under a condition similar to the above predetermined condition as a first auxiliary sample on the basis of the past event occurrence data 143 accepted as the input in step S101. In this example, the similar condition indicates, for example, a condition such as a period of time between 10:00 and 11:00 on weekdays in August and October before and after September as described above.
In step S104, for example, the CPU 11 creates a plurality of allocation plans of ambulances for a sick/injured person occurrence event on the basis of the above ambulance data 141 in
In step S105, the CPU 11 applies the first main sample output in step S102 and the first auxiliary sample output in step S103 to each of the plurality of allocation plans created in step S104 and evaluates whether or not each of the plurality of allocation plans satisfies the evaluation criteria. Then, among the plurality of allocation plans, the CPU 11 stores effective allocation plans that satisfy the evaluation criteria in the storage 14 as, for example, the above effective allocation plans 144 in
In step S111 of
In step S112, the CPU 11 extracts one piece of the event occurrence data.
In step S113, the CPU 11 dispatches the closest ambulance for the event occurrence data extracted in step S112 and gives, to the dispatched ambulance, a dispatch impossible flag indicating that the dispatched ambulance cannot be allowed for the next dispatch for a certain period of time.
In step S114, the CPU 11 calculates a distance from a fire station where the ambulance to which the dispatch impossible flag is given in step S113 is allocated to a site at which the ambulance arrives, stores the calculated distance in the storage 14, returns to step S111, and repeats the processing for all pieces of the event occurrence data.
Meanwhile, in step S115, the CPU 11 calculates an average distance required to arrive at the site, and returns to step S105 in
Next, processing in phase 2 will be described. The processing in phase 2 is executed, for example, at a fixed time before 10:00 (e.g. 9:00) on weekdays in September. The second sample output unit 104 obtains, for example, an occurrence frequency of a sick/injured person in each area between 10:00 and 11:00 on weekdays in the last month on the basis of the latest event occurrence data 145 and samples future pseudo occurrence data on the basis of a result thereof. The latest event occurrence data 145 indicates the latest past sick/injured person occurrence data accumulated by immediately before. The number of days to be sampled is, for example, 100 days. In addition, as in the processing in phase 1 described above, the second main sample output unit 104A and the second auxiliary sample output unit 104B may share roles, increase or decrease the occurrence frequency by a certain value, and output samples, or may generate samples on the basis of the random variable by using MCMC described above in
The second allocation evaluation unit 105 evaluates all the effective allocation plans 144 accumulated in the storage 14. The evaluation method herein is similar to the evaluation method in the processing in phase 1 described above, except that the acquired allocation plans are not the allocation plans created by the allocation plan creation unit 102, but are the effective allocation plans 144 evaluated by the first allocation evaluation unit 103.
In step S121 of
In step S122, the CPU 11 outputs future event occurrence data that may occur under a predetermined condition as a second main sample on the basis of the latest event occurrence data 145 accepted as the input in step S121. In this example, the predetermined condition indicates, for example, a condition such as a period of time between 10:00 and 11:00 on weekdays in September as in phase 1 described above.
In step S123, the CPU 11 outputs future event occurrence data that may occur under a condition similar to the above predetermined condition as a second auxiliary sample on the basis of the latest event occurrence data 145 accepted as the input in step S121. In this example, the similar condition indicates, for example, a condition such as a period of time between 10:00 and 11:00 on weekdays in August and October before and after September as in phase 1 described above.
In step S124, the CPU 11 applies the second main sample output in step S122 and the second auxiliary sample output in step S123 to each of the effective allocation plans 144 (see
In step S125, the CPU 11 outputs a final evaluation result obtained by the reevaluation in step S124, and the processing in phase 2 by the allocation search program ends.
In this way, the effective allocation plans are evaluated again on the basis of the latest occurrence data rows. As a result, an evaluation result may be different from the above evaluation result in
The processing in phase 2 can be performed by a general-purpose PC in about several tens of seconds to several minutes. Therefore, for example, it is possible to quickly find and employ appropriate allocation of the resources on the basis of the latest event occurrence data on the day when the allocation of the resources is desired to be changed.
The above method can be applied to other cases. For example, in case that the number of ambulance squads is reduced for some reason, it is also possible to obtain effective allocation with a small number of ambulance squads in advance and use the allocation.
As described above, according to this embodiment, in a case where it is necessary to determine effective allocation of resources in a relatively short time, it is possible to obtain robust allocation of the resources that can expect an effect even if a prediction result of occurrence of an event is not completely right.
Allocation search processing that is executed by the CPU reading software (program) in the above embodiment may be executed by various processors other than the CPU. Examples of the processors in this case include a programmable logic device (PLD) whose circuit configuration can be changed after manufacturing, such as a field-programmable gate array (FPGA), and a dedicated electric circuit that is a processor having a circuit configuration exclusively designed for executing specific processing, such as an application specific integrated circuit (ASIC). Further, the allocation search processing may be executed by one of the various processors or may be executed by a combination of two or more processors of the same type or different types (e.g. a combination of a plurality of FPGAs or a combination of a CPU and an FPGA). Furthermore, a hardware structure of the various processors is, more specifically, an electric circuit in which circuit elements such as semiconductor elements are combined.
In the above embodiment, the aspect in which the allocation search program is stored (installed) in advance in the storage has been described, but the embodiment is not limited thereto. The program may be provided by being stored in a non-transitory storage medium such as a compact disk read only memory (CD-ROM), a digital versatile disk read only memory (DVD-ROM), and a universal serial bus (USB) memory. The program may be downloaded from an external device via a network.
Regarding the above embodiment, the following supplementary notes are further disclosed.
An allocation search device comprising:
A non-transitory storage medium storing a program executable by a computer to execute allocation search processing,
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2020/040462 | 10/28/2020 | WO |
