JP5762873B2 - Repeating schedule generation apparatus and method - Google Patents

Description

  The present invention relates to an iterative schedule generation apparatus and method capable of generating a regular and unbiased schedule pattern in scheduling using a network.

  For example, in the scheduling of radiation therapy, it is important to generate a schedule that can be easily recognized by medical staff in order to perform treatment safely. Like regular schedules that repeat on a weekly basis, it is easier for medical staff to recognize if the day of the week is regular.

  However, radiotherapy may include a protocol in which treatment is performed only once every 2 to 3 weeks depending on the number and ratio of protocols (morbidity types). For this reason, in scheduling, a schedule that is not necessarily repeated for one week but is repeated in units of a plurality of weeks is created. This method of putting a schedule for a plurality of weeks or the result of putting it is called a schedule pattern.

  In order to provide such a schedule pattern with regularity that is easy for people to remember, we want to make the previous week and the next week similar. On the other hand, in order to improve the usability of the schedule pattern, it is desirable that the schedule is not biased without scheduling a specific protocol on a specific day.

  However, it is difficult to generate a regular and unbiased schedule pattern. Although this situation is scheduling other than radiotherapy, two prior arts will be described as examples (see, for example, Patent Documents 1 and 2).

  In the first prior art, when the amount of work aggregated in units of months exceeds the upper limit value, shift processing is performed in the previous month or the next month in the order of work having the lowest evaluation value SW among the work items assigned to the excess month. A schedule without excess work is created.

  In the first prior art, in creating the work schedule for each unit period included in a fixed plan period under the inspection cycle set for each of the plurality of facilities, the optimization has been made without excessive or wasteful optimization. A scheduling method is provided. In this scheduling method, it is possible to make a daily plan. However, since the work performed by each worker is different every day, it is necessary for the worker to check the work item every time and perform the work.

  However, the necessity of confirming the work item every time places a burden on the worker. If the day's work is decided according to the day of the week, it is very easy to remember, but the work is not always entered in a weekly unit.

  Thus, in the first prior art, it is difficult to create a schedule having regularity that is easy for an operator to recognize. Further, in the first prior art, it is difficult to easily generate a schedule that satisfies such a restriction when a worker having a specific skill can work only on a specific day of the week.

  Next, the second prior art includes a step of assigning each periodic work according to each work interval to a work point set for each unit work interval in the work cycle. In the assignment step, each periodic work is assigned in order. In the assignment step, for each periodic work to be assigned, (1) the periodic work is provisionally assigned to each of all assignment patterns, and “assigned periodic work for each work point is assigned to each provisionally assigned pattern. The standard deviation of the “total work time of the work group” is calculated, and (2) the regular work is assigned to the assignment pattern that minimizes the standard deviation.

  In such a second prior art, by determining the work cycle for creating the periodic work plan and optimizing the allocation pattern, the worker repeats the same tool change work, thereby improving the skill of the worker. This is easy and reduces the work burden on the workers.

  However, in practice, for example, when a work that is executed once for every even number of work points and a work that is executed once for every odd or prime number of work points are mixed, the combination of replacements at each work point is Since it becomes equal to the work cycle for creating the periodic work plan, it is difficult to create a schedule for a worker having a specific skill to repeatedly perform a specific work.

  Therefore, it is difficult to apply the second prior art to the scheduling of radiation therapy related to a specific skill or work such as making a fixture. Furthermore, in the second prior art, the same worker can only be assigned, for example, even-numbered or odd-numbered, and work points handled by the same worker cannot be given freely.

JP-A-9-120417 JP 2007-102684 A

  As described above, in the first prior art, it is difficult to create a schedule having regularity that is easy for a worker to recognize when work is not necessarily performed in units of one week.

  Further, in the second prior art, even if the work point assigned to the same worker is freely given, a specific work can be performed with regularity while leveling the work amount at the work point so as not to be biased. It is difficult to create a schedule that is easy to gather.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an iterative schedule generation apparatus and method capable of generating a regular and unbiased schedule pattern.

  One aspect of the present invention assigns each protocol to indicate the date of the base treatment process for each radiotherapy protocol for each combination of day of the week and week in the repeating period that is repeated periodically. A repetitive schedule generation device for generating a schedule pattern, wherein each protocol showing the treatment site and the number of irradiation treatment steps in each treatment step, and allocation work information defining the number of patients of each protocol A first storage means for storing; a second storage means for storing treatment process information defining each treatment process for each protocol and a treatment process serving as a base point of each treatment process; and between each treatment process A third storage means for storing restriction information defining restrictions on the number of days to be freed up and the number of times of the irradiation treatment process; A fourth storage means for storing allocation time information individually indicating a combination of a day of the week and a week that can be allocated, and a day type individually allocated to a combination of a day of the week and a week in the allocation time information Fifth storage means for storing allocation time type information, and a plurality of allocation time nodes individually indicating combinations of days of the week and weeks indicated in the allocation time information based on the allocation time information and the allocation work information; , A plurality of protocol nodes individually indicating each protocol defined in the allocation work information, a plurality of allocation arcs individually indicating a combination of each allocation time node and each protocol node, and the number of patients of each protocol Allocation network creation means for creating an allocation network including the allocation work information and the allocation time information. A flow rate limit value calculating means for calculating a flow rate limit value for limiting the flow rate as a possible number of patients, and a plurality of allocated time types individually indicating the day type based on the allocated time information and the allocated time type information Type element generating means for generating a node and a plurality of allocation time type arcs individually indicating valid allocation time type nodes of the allocation time type nodes, the allocation work information, the allocation time information, and the allocation Based on the time type information, for each protocol, the used branch number limit value calculating means for calculating the used branch number limit value for limiting the number of used branches of each allocated time type arc, the allocation network, the flow rate limit By integrating the value, each allocation time type node, each allocation time type arc, and the use branch number limit value, each allocation time node and each protocol A limit network including a node, a number of patients, the number of patients, the flow rate limit value, each of the allocated time type nodes, each of the allocated time type arcs, and the used branch number limit value. Based on the restriction network creation means, the restriction network, the treatment process information, and the restriction information, the schedule pattern is generated, and the schedule pattern generation means for writing the generated schedule pattern in the schedule pattern storage means; It is a repetitive schedule generation device comprising: the schedule pattern storage unit that stores the schedule pattern that is loaded; and the schedule pattern output unit that reads the schedule pattern from the schedule pattern storage unit and outputs the read schedule pattern.

  As described above, according to the present invention, the flow rate limit value limits the number of patients assigned to a specific day of the week and week combination and limits the number of days of the week according to the use branch number limit value. It is possible to generate a regular and uniform schedule pattern.

It is a block diagram which shows the structure of the repetition type schedule production | generation apparatus which concerns on 1st Embodiment. It is a schematic diagram which shows an example of the schedule pattern in the embodiment. It is a schematic diagram which shows an example of the allocation work information in the same embodiment. It is a schematic diagram which shows an example of the allocation time information in the embodiment. It is a schematic diagram which shows an example of the allocation time kind information in the same embodiment. It is a schematic diagram which shows an example of the planning period information in the embodiment. It is a schematic diagram which shows an example of the protocol information in the embodiment. It is a schematic diagram which shows an example of the treatment process information in the embodiment. It is a schematic diagram which shows an example of the constraint information in the embodiment. It is a schematic diagram which shows an example of a holiday treatment day pattern information in the embodiment. It is a schematic diagram which shows an example of the days type information in the embodiment. It is a schematic diagram which shows an example of the allocation network in the embodiment. It is a schematic diagram which shows the other example of the allocation network in the embodiment. It is a schematic diagram which shows an example of the restriction | limiting network in the embodiment. It is a schematic diagram which shows the other example of the restriction | limiting network in the same embodiment. It is a flowchart for demonstrating the operation | movement in the embodiment. It is a schematic diagram which shows an example of the treatment process of the radiotherapy in the embodiment. It is a schematic diagram for demonstrating the combination of the number of days type and rest treatment day pattern in the embodiment. It is a schematic diagram which shows the example of calculation of the flow volume upper limit in the same embodiment. It is a schematic diagram for demonstrating calculation of the use branch number limit value in the embodiment. It is a schematic diagram which shows the example of calculation of the use branch number upper limit in the same embodiment. 5 is a flowchart for explaining in detail an allocation network creation process in the embodiment. It is a schematic diagram which shows the example of calculation of the multidimensional cost in the same embodiment. It is a schematic diagram for demonstrating the effect in the same embodiment. It is a schematic diagram for demonstrating the effect in the same embodiment. It is a block diagram which shows the structure of the repetition type schedule production | generation apparatus which concerns on 2nd Embodiment. It is a schematic diagram which shows the example of a display of the limiting value display part in the embodiment. It is a schematic diagram which shows the example of a display of the schedule pattern display part in the embodiment. It is a schematic diagram which shows the example of a setting when the restrictions in the embodiment are loosened. It is a schematic diagram which shows the example of a display of the schedule pattern display part in the embodiment. It is a schematic diagram which shows the example of a setting when the restrictions in the embodiment are tightened. It is a schematic diagram which shows the example of a display of the schedule pattern display part when the restrictions in the embodiment are severe. It is a schematic diagram which shows the example of the schedule pattern in 3rd Embodiment. It is a schematic diagram which shows the example of the allocation network in the embodiment. It is a schematic diagram which shows the example of the restriction | limiting network in the embodiment. It is a schematic diagram which shows the example of calculation of the use branch number upper limit in the same embodiment. It is a schematic diagram which shows the example of a change of the restriction | limiting network in the same embodiment. It is a schematic diagram which shows the example of calculation after the change of the use branch number upper limit in the embodiment. It is a schematic diagram which shows the example of the schedule pattern after the change in the same embodiment.

  Each embodiment will be described below with reference to the drawings, but before that, an outline of each embodiment will be described.

  In the first and second embodiments, the radiotherapy scheduling problem is regarded as a network flow problem. For example, a special node and an arc of an allocation time type are added to the allocation network, and this arc is set for each work type. Regularity is given to the generated schedule pattern by the configuration that limits the number of lines to be used. The number limit can be loosely applied only at the upper limit, or can be strictly restricted by providing upper and lower limits, and can be easily adjusted. Since the restriction on the number of branches used can be formulated as a restriction expression for a mixed integer linear programming problem, optimization is possible.

  In addition, the flow limit value that restricts the flow rate through the assigned arc that assigns the assigned work to the assigned time is also described as a constraint, thereby enabling optimization to optimize the objective function (for example, leveling).

  An example of how to calculate the upper and lower limit values used for constraints is shown in the first and second embodiments.

  In the first and second embodiments, an example of the MILP (mixed integer linear programming) formulation of the optimization problem in radiation therapy scheduling is described.

  In the third embodiment, as a modified example of the first embodiment, when the grouping of the assigned time type is divided for each worker, each worker can generate a schedule pattern for repeatedly performing the work of a specific work type. Says. In the example described in the third embodiment, there are changes such as using a work point type node instead of the assigned time type node and using a work point node instead of the assigned time node. In the third embodiment, the name of information and the like are changed from those in the first embodiment, but the contents of processing for processing or calculating information are the same as those in the first embodiment.

  The above is the outline of each embodiment. Subsequently, each embodiment will be specifically described.

<First Embodiment>
FIG. 1 is a block diagram illustrating a configuration of a repetitive schedule generation device according to the first embodiment, and FIG. 2 is a schematic diagram illustrating an example of a schedule pattern generated by the repetitive schedule generation device. The repetitive schedule generation device generates a schedule pattern of a repetitive period that is periodically repeated by the processing of the units 1 to 12.

  The schedule pattern in the repetitive scheduling means an assignment work and an assignment time, and a predetermined number of each assignment work is assigned to any of the assignment times.

  In this type of schedule pattern, for example, as shown in FIG. 2, for each combination of the day of the week and week (assigned time) during the repetition period, a treatment that is a base point for each radiotherapy protocol (morbidity type) Each protocol is assigned to indicate the date of the process (assignment work). In FIG. 2, “Month 1” is an abbreviation for “Monday of the first week”. “Lung 1” is an abbreviation that represents a protocol indicating that the treatment site is “lung” and the radiation treatment process is “once”. Similarly, “liver 8” and “front 16” are abbreviations representing protocols indicating that the treatment sites are “liver” and “prostate”, respectively, and the irradiation treatment process is “8 times” and “16 times”. It is. These abbreviations are used in the following description. The schedule pattern includes each protocol by the number of patients of each protocol. The assignment of 10 “liver 8” to the schedule pattern corresponds to the number of patients of the protocol “liver 8” being 10.

  On the other hand, as shown in FIG. 1, the repetitive schedule generation device includes an information input unit 1, an allocation work information storage unit 2, an allocation time information storage unit 3, an allocation time type information storage unit 4, a schedule information storage unit 5, an allocation A network creation unit 6, a flow rate limit value calculation unit 7, a type element generation unit 8, a use branch number limit value calculation unit 9, a limit network creation unit 10, a schedule pattern generation unit 11, and a schedule pattern storage and output unit 12 are provided. .

  The information input unit 1 is an input interface between the user and each of the units 2 to 12, and has a function of writing various information to the storage units 2 to 5 according to user operations, for example.

  Each of the storage units 2 to 5 is a storage device or a storage area of the storage device that can be read / written from each of the units 1 and 6 to 12. For example, the information written from the information input unit 1 can be read from the units 6 to 12. Store readable. Further, for example, the storage units 2 to 5 may store data in the middle of processing of the units 6 to 11.

Specifically, as shown in FIG. 3, the allocation work information storage unit (first storage unit) 2 includes each protocol j indicating the treatment site and the number of irradiation treatment steps in each treatment step, and each protocol j. storing allocation task information for defining the patient number d _j.

  As shown in FIG. 4, the allocation time information storage unit (fourth storage unit) 3 stores allocation time information i that individually indicates combinations of days of the week and weeks to which each protocol j can be allocated.

  As shown in FIG. 5, the allocated time type information storage unit (fifth storage means) 4 is an allocated time type that specifies the day type w that is individually allocated to the combination of the day of the week and the week in the allocated time information i. Store information.

  As shown in FIGS. 6 to 11, the schedule information storage unit (second, sixth to eighth storage units) 5 includes planning period information, protocol information, treatment process information, and constraints written from the information input unit 1. Information, rest treatment day pattern information and number of days type information are stored. Of the information in the schedule information storage unit 5, for example, planning period information, protocol information, holiday treatment day pattern information, and number of days type information are arbitrary additional items from the viewpoint of creating a schedule pattern. Can be omitted. However, it is preferable not to omit the planning period information, the holiday treatment day pattern information, and the day type information from the viewpoint of formulating a mixed integer linear programming problem described later. For convenience of explanation, each piece of information is collectively stored in the schedule information storage unit 5, but may be modified so as to be stored in another storage unit as appropriate.

  The planning period information is information indicating the planning period T as shown in FIG.

  The protocol information is information defining each protocol j and the number of irradiation treatment steps as shown in FIG.

  As shown in FIG. 8, the treatment process information is information defining each treatment process s. For example, each treatment process s for each protocol j and a treatment process (hereinafter referred to as a base point) among the treatment processes s. It is also information that defines the standard treatment process. In the standard treatment process, an upper limit of the working time for performing the standard treatment process on each day can be specified. It is stipulated that the schedule for other treatment processes is generated at a relative position (number of days in between) from the reference treatment process. The reference treatment process is not limited to the creation of a fixture as the first treatment process, and CT simulation and first irradiation can be designated.

  As shown in FIG. 9, the constraint information is information defining treatment process constraints, specifically, information defining constraints on the number of days and the number of irradiation treatment steps between the treatment steps.

  As shown in FIG. 10, the rest treatment day pattern information is information defining a pattern k of a day (in the figure, a rest treatment day of irradiation) on which the irradiation treatment process out of each treatment process is rested.

  As shown in FIG. 11, the number-of-days information is information that defines the number of days l that are available between treatment steps.

  The allocation network creation unit 6 has a function of creating an allocation network based on allocation time information and allocation work information in the storage units 2 and 3.

As shown in FIG. 12, the allocation network includes a plurality of allocation time nodes “Month 1”,..., “Fri” 2 each indicating a combination of a day of the week and a week indicated by the allocation time information i, and the allocation work. A plurality of protocol nodes “lung 1”,..., “Previous 16” individually indicating each protocol j defined in the information, and a plurality of allocation arcs α individually indicating a combination of each allocation time node and each protocol node. , And the number of patients d _{j for} each protocol j. As shown in FIG. 13, the allocation network may further include a holiday treatment day pattern node that individually indicates a holiday treatment day pattern k, and a number of days type node that individually indicates the number of days type l that is in between. In the present embodiment, the case of the allocation network shown in FIG. 13 will be mainly described as an example.

The flow rate limit value calculation unit 7 limits the flow rate as the assignable number of patients d _j for each assigned arc α based on the assigned work information and the assigned time information i in the storage units 3 and 5. It has a flow rate limit value calculation function that calculates the value. Note that the flow rate limit value is the upper limit value or the upper limit of the number of the same protocol on the same day (same column) in FIG. 2 (eg, the number of “liver 8” on the day of “Thu 1” is “2”). It can also be interpreted as the lower limit value.

Here, the flow restriction value calculation function, represents each protocol specified in assigned work information and j represents the number of patients in each protocol j of the day during the repetition period and d _j, each shown in allocated time information i When a set of days is expressed as I, the number of each day included in the set I is expressed as | I |, and a flow rate upper limit value among the flow rate limit values is expressed as FLOW (j), the following formula (1) As described above, a flow rate upper limit calculation function for calculating the flow rate upper limit value FLOW (j) as the flow rate limit value may be provided.

  The type element generation unit 8 has a plurality of allocation time type nodes individually indicating the day type w as shown in a part of FIG. 14 based on the allocation time information i and the allocation time type information in the storage units 3 and 4. A function for generating “month”,..., “Gold” and a plurality of assigned time type arcs β individually indicating valid assigned time type nodes of each assigned time type node “month”,. Have The term “type element” may be read as “allocation time type node and allocation time type arc”.

  The use branch number limit value calculation unit 9 uses the use branch number limit value for limiting the use branch number of each allocation time type arc for each protocol j based on the allocation work information, the allocation time information i, and the allocation time type information. It has a function to calculate the number of used branches limit value. Note that the number of used branches is limited to the number of valid day types (month, Tuesday, Thursday, Friday) for the combination of day of the week and week “Month 1”,. Can be interpreted as an upper limit value or an upper and lower limit value of “4”).

Here, the use branch number limit value calculation function represents the number of types of weeks indicated in the allocation time information i as h, and the use branch number upper limit value of the use branch number limit values as USE (j). As shown in the following formula (2), a use branch number upper limit value calculation function for calculating the use branch number upper limit value USE (j) as the use branch number limit value may be provided.

The restriction network creation unit 10 has a function of creating a restriction network by integrating the assignment network, the flow restriction value, each assignment time type node, each assignment time kind arc, and the number of used branches. As shown in FIG. 14, the restriction network includes each allocation time node “month 1”,..., “Gold 2”, each protocol node “lung 1”,..., “Front 16”, each allocation arc α, Number of patients _dj , flow rate limit value (not shown), each assigned time type node “month”,..., “Gold”, each assigned time type arc β, and used branch number limit value (not shown) Including. As shown in FIG. 15, the restriction network may further include a holiday treatment day pattern node that individually indicates a holiday treatment day pattern k, and a number of days type node that individually indicates the number of days l that are in between. In the present embodiment, the case of the restricted network shown in FIG. 15 will be mainly described as an example. In addition, the term “restricted network” may be read as “network with assigned time type”.

  The schedule pattern generation unit 11 is a schedule pattern generation function that generates a schedule pattern based on the restriction network, treatment process information, and constraint information, and a schedule pattern document that writes the generated schedule pattern to the schedule pattern storage and output unit 12 With a built-in function.

Here, the schedule pattern generation function represents a day in the planning period T as {0,..., T}, represents each treatment process as s, represents a set of the treatment processes s as S, and is included in the set I Represents a set of patterns k as K, a set of protocols j as J, a set of number of days l as L, a day of the week as w, and the day w Is represented as W, and the set of days of the day of the week w in the repetition period is represented as DAY (w), (w∈W). In the combination ikjl of each day i, pattern k, each protocol j, and number of days l the number of patients is expressed as _y ikjl, the upper bound of the possible values of the number of patients y _Ikjl expressed as gamma (where, r = max {FLOW (j ) | j∈J}),

The upper limit value that can be taken is represented by υ (where υ = max {USE (j) | j∈J}), and the working time of the treatment step s on day t when one patient is assigned in the combination ikjl. cost showing expressed as _c 'tsikjl, a 0-1 variable to 1 when assigned by myself the patient in combination Ikjl expressed as x _ikjl, a case assigned the protocol j to week w to 1 0-1 variable _Is expressed as u _wj , the whole integer is expressed as Z, and the evaluation value of the mixed integer linear programming problem shown in the following formula (3) is expressed as z, the mixed integer linear programming problem is solved so as to minimize the evaluation value z Thus, a solver function for generating a schedule pattern may be provided.

  The schedule pattern storage and output unit 12 includes a schedule pattern storage unit (not shown) that stores the schedule pattern written from the schedule pattern generation unit 11, reads the schedule pattern from the schedule pattern storage unit, and reads the schedule pattern It has a function to output the schedule pattern.

  Next, the operation of the repetitive schedule generation device configured as described above will be described with reference to FIG. The following description will be given by taking as an example the case of creating a 2-week repetition schedule in heavy ion radiotherapy.

  First, an outline of scheduling in heavy particle beams will be described. Scheduling here is to create a schedule frame for the planning period from the frequency of visiting patients and the ratio of the patient's protocol (condition type).

  In hospitals, in order to avoid exceeding daily loads, for example, a semi-scheduled schedule frame (treatment frame) is created in advance, and a method is often used in which the frame is filled when a new disease is reserved. The treatment window is also useful when making appointments several months ahead in combination with other therapies or when estimating the number of patients coming to the hospital. The schedule is created and filled by satisfying both the regularity that is easy to understand for the staff and the similar schedule repeated every week or every two weeks and the efficiency of load leveling. It aims to treat more patients safely.

  In the heavy ion radiotherapy, there are the treatment steps shown in FIG. 17, and among these, there are three treatment steps for scheduling, for example, creation of a fixture, CT (computerized tomography) simulation, and irradiation. In addition, a process such as rehearsal may be included in the scheduling. The irradiation treatment process is performed a plurality of times, and the number of times varies depending on the type of protocol.

  Regarding a treatment process for performing scheduling, a treatment date (schedule) for one patient is defined as one schedule frame (hereinafter referred to as a frame). The set of all patient frames for half a term is referred to herein as a treatment window. Each slot has a specific protocol, but it is not determined which patient it is assigned to.

  The frame creation problem in heavy ion radiotherapy (hereinafter referred to as frame creation problem) is the conditions that can be executed for each treatment process in each protocol according to the size of the facility (or the number of hospital staff), such as the number of rooms, and various schedules. When an optimization objective is given, it is to obtain a set of frames (scheduled treatment frames) in each patient that satisfies the feasible conditions and improves the optimization objective.

  From the viewpoint of safety and convenience, it is preferable that the treatment window has regularity on a weekly basis. In practice, there are protocols less than once a week, so the repetition period is set according to the protocol with the lowest frequency, and the schedule of this repetition period is repeated periodically during the planning period. Further, regarding the schedule during the repetition period, the regularity in units of weeks and the regularity in which a protocol with a large number of people is allocated as evenly as possible on each day are emphasized.

  On the other hand, in order to reduce the burden on the staff, it is desirable that the working time of each day be leveled as much as possible. Since irregular cases occur on holidays and equipment maintenance, working hours must be considered throughout the planning period. Minimize working hours on the day with the most work time for irradiation during the planning period. Alternatively, the total work time on the day having the maximum total work time may be minimized among the total work hours obtained by summing up the work hours of the respective treatment steps assigned every day during the repetition period. .

  Since the number of days left between the processes is given, the work time for irradiation on each day (during the planning period) when one person is assigned to start treatment (preparation of fixtures) for each protocol on each day during the repetition period. Can be calculated.

  The treatment schedule window can be generated by determining a schedule pattern including designation of the start date of each protocol during the repetition period.

  In performing such scheduling, the information input unit 1 assigns the allocation work information, the allocation time information i, and the allocation time type information to the allocation work information storage unit 2, the allocation time information storage unit 3, and the allocation according to user operations. Write to the time type information storage unit 4.

  Similarly, the information input unit 1 writes planning period information, protocol information, treatment process information, constraint information, holiday treatment day pattern information, and number of days type information in the schedule information storage unit 5 in accordance with a user operation.

Subsequently, the allocation network creation unit 6 is based on the allocation time information i and the allocation work information in each of the storage units 2 and 3, and includes a plurality of individually indicating combinations of days of the week and weeks indicated by the allocation time information i. Allocation time nodes “Month 1”,..., “Friday 2” and a plurality of protocol nodes “lung 1”,..., “Previous 16” individually indicating each protocol j defined in the allocation work information, and each allocation time and multiple assignments arc α indicating individually combined with each protocol node and the node, creating an allocation network including a number of patients d _j of each protocol j (S11). Details of step S11 will be described later.

  For example, as shown in FIG. 12, the allocation network represents which allocation work is allocated to which allocation time as a bipartite graph matching, and an allocation arc α that is an arc between the allocation time and the allocation work. A set of flow rates flowing through the flow represents one schedule pattern element.

  As shown in FIG. 13, the allocation network may be a network in which options for the number of days l that are spaced between each of the allocation time and each allocation operation and the treatment day pattern k are expanded.

  The number of days available in the choice of l is a solution to the problem that the work days of treatment processes other than the standard treatment process cannot be uniquely determined. It is what. The choice of the non-treatment pattern k is optimal from multiple patterns of irradiation days when there are more weeks in the week than the maximum number of irradiations per week per patient. This makes it possible to select a good pattern.

  Further, in addition to making it possible to select two points of the number of days type l and the rest treatment day pattern k, as shown in FIG. 18, there is an interval between treatment steps for each patient of the same protocol j on the same start date i. The combination of the number of days type l and the holiday treatment day pattern k is made one type (only one assigned arc α between the node k and the node l is used). This solves the problem that the work days of treatment processes other than the reference treatment process cannot be uniquely determined.

  The flow rate limit value calculation unit 7 calculates a flow rate limit value for each set (α, p) of each assigned arc α and each assigned work p in the assigned network (S12, S13). Each assignment operation p corresponds to each protocol j.

  The flow rate restriction is such that the assignment is not biased toward a specific assignment arc α. For example, the upper limit value can be calculated as follows.

That is, the flow rate upper limit value FLOW (j) indicates the upper limit that may be assigned per day during the repetitive period of the protocol j. Therefore, when the number of elements of the set I is | I | Based on this, the number of patients d _j of protocol j is divided by the number of days | I | Note that “the smallest integer greater than or equal to the division result” may be read as “the smallest integer that does not fall below the division result” or “the integer that is rounded up after the decimal point of the division result”. Here, FIG. 19 shows a calculation example of the flow rate upper limit value FLOW (j) based on the information shown in FIGS. For example, when the protocol j is “liver 8”, d _j = 10, | I | = 8, and based on the formula (1), FLOW (j) is d _j / | I | (= 10/8) or more Is calculated as the smallest integer “2”.

  Next, the type element generation unit 8 assigns the allocation time type nodes “month”,..., “Gold” as shown in FIG. 15 based on the allocation time information and the allocation time type information in the storage units 3 and 4. Are generated (S14), and an arc is generated between each allocated time type node and source node, and between each allocated time type node and the allocated time node corresponding to the allocated time type (S15, S16).

  The used branch number limit value calculation unit 9 calculates the used branch number limit value for each set (β, p) of each allocation time type arc β (FIG. 20) and each allocation work p that has been generated (S 17, S 18). . The use branch number limit value limits the number of types of allocation time to which the allocation work p is allocated (the number of allocation time type arcs β). For example, the upper limit value can be calculated as follows.

That is, the upper limit value USE (j) for use branches is set to h as the number of weeks in the repetition period, and the number of patients d _j of the protocol j is set to the number h of weeks in the repetition period based on the above-described equation (2). It can be calculated as the smallest integer greater than the result of division. FIG. 21 shows a calculation example of the used branch number upper limit value USE (j) based on the information shown in FIGS. For example, when protocol j is “liver 8”, d _j = 10 and h = 2, and USE (j) is the smallest integer equal to or greater than d _j / h (= 10/2) based on equation (2). Although it is calculated as “5”, since the maximum number of branches used for the allocation time type arc β is 4, the calculation result is “4”.

  The restricted network creation unit 10 obtains the assigned network, the flow upper limit value FLOW (j), each assigned time type node “month”,..., “Fri”, each assigned time type arc β, and the use branch number upper limit value USE (j). By integrating, a restricted network as shown in FIG. 15 is created (S19).

  This restricted network can be formulated as a mixed integer linear programming problem as shown in Equation (3) above, and is a general-purpose solver that solves mixed integer linear programming problems such as CPLEX, exact solution, heuristic solution, metaheuristic solution , Can be optimized.

  In the formulation, symbols related to input data are defined below.

{0, ..., T}: Day during the planning period.

S: A set of treatment steps.

I: A set of days during the repeating period.

K: A set of irradiation patterns.

J: Protocol set.

L: A set of days left in between.

W: A set of days of the week.

d _j : Number of patients in protocol j on the day during the repeat period.

Next, symbols related to values calculated from input data are defined.
DAY (w), (w∈W): A function that returns a set of days in the repetition period where the day of the week is w.

USE (j), (j∈J): A function that returns the upper limit of the number of types of days that protocol j can take.

FLOW (j), (j∈J): A function that returns an upper limit amount that can be allocated per day of the day in which the protocol j is repeated.

γ: Upper bound of the value that y _ikjl can take.

c ′ _tsikjl : Cost of treatment step s on day t when one patient is assigned to i, k, j, l. Symbols relating to variables being formulated are defined below.

y _ikjl : An integer variable indicating the number of patients assigned to i, k, j, l.

x _ikjl : A 0-1 variable that takes 1 when at least one patient is assigned to i, k, j, l.

u _wj : 0-1 variable that takes 1 when protocol j is assigned to day w.

  However, Z is a symbol indicating the whole integer.

  Using the above definition, the restricted network is formulated as a mixed integer linear programming problem as shown in Equation (3).

  The schedule pattern generation unit 11 minimizes the evaluation value z of the above-described formula (3) based on the restriction network that can be formulated in this way, the treatment process information and the constraint information in the schedule storage unit 5. By solving the mixed integer linear programming problem, a schedule pattern as shown in FIG. 2 is generated (S20).

  Thereafter, the schedule pattern generation unit 11 writes the generated schedule pattern in the schedule pattern storage and output unit 12.

  The schedule pattern storage and output unit 12 stores a written schedule pattern in a schedule pattern storage unit (not shown), reads the schedule pattern from the schedule pattern storage unit, and outputs the read schedule pattern.

  Next, the allocation network creation processing in step S11 described above will be described in detail.

  In the example of radiotherapy, the assigned time indicates the start date of a treatment process that serves as a reference, for example, the creation of a fixture, and the assigned work indicates the type of protocol. The allocation network shown in FIG. 13 is based on the allocation work information in the allocation work information storage unit 2, the allocation time information i in the allocation time information storage unit 3, and each information in the schedule information storage unit 5. Created by the allocation network creation unit 6. Such allocation network creation processing will be described in detail with reference to FIG.

  First, the allocation network creation unit 6 creates dummy nodes one by one, a sink node and a source node (S11-1).

  The allocation network creation unit 6 generates allocation work nodes “lung 1”,..., “Front 16” of each protocol j based on the allocation work information in the allocation work information storage unit 2 (S11-2).

  Subsequently, the assignment network creation unit 6 repeats the following steps S11-4 to S11-6 for all assignment work nodes “lung 1”,..., “Previous 16” (S11-3).

  That is, the allocation network creation unit 6 generates an arc between the allocation work node and the sink node for each allocation work node “lung 1”,..., “Previous 16” (S11-4). Based on the number-of-days-type information, the number-of-days-type nodes for the number of types are generated (S11-5). Further, the allocation network creation unit 6 generates an arc between the day type node and the allocation work node (S11-6).

  After the end of step S11-3 which repeats steps S11-4 to S11-6, the allocation network creation unit 6 determines the allocation time node “month 1”,... Based on the allocation time information i in the allocation time information storage unit 3. , “Gold 2” is generated (S11-7).

  Thereafter, as described above, the following steps S11-9 to S11-11 are repeated for all allocation time nodes “month 1”,..., “Gold 2” (S11-8).

  That is, the allocation network creation unit 6 generates an arc between the allocation time node and the source node for each allocation time node “Month 1”,..., “Friday 2” (S11-9). Based on the rest treatment day pattern information, a rest treatment day pattern node is generated (S11-10). Moreover, the allocation network creation part 6 produces | generates the arc between a holiday treatment day pattern node and an allocation time node (S11-11).

  After step S11-8, in which steps S11-9 to S11-11 are repeated, assignment network creation unit 6 generates assignment arc α (S11-12). In the allocation network shown in FIG. 13, the generation process of the allocation arc α may be performed by forming a complete bipartite graph with the number of days type nodes that are available for each allocation operation and the non-medical treatment day pattern node for each allocation time. When there are no rest treatment day pattern nodes and number of days type nodes that are left between them as in the assignment network shown in FIG. 12, each assignment work node “lung 1”,..., “Front 16” and each assignment time node “ A complete bipartite graph may be composed of “month 1”,..., “Gold 2”. An arc generated so as to form a complete bipartite graph with a node related to allocation work and a node related to allocation time is called an allocation arc α.

  The allocation network creation unit 6 calculates a multidimensional cost for each allocation arc α based on each information in the schedule storage unit 5 as shown in FIG. 23 (S11-13, S11-14), and the allocation network Create

  Here, the multidimensional cost calculation method in step S11-13 will be described in detail.

The cost c ' _{tsikjl on} the day t during the planning period when the patient is assigned to the day i during the repetition period, the protocol j, the treatment treatment day pattern k, and the number of days l in between is calculated as follows: it can.

First, for a day i in the repetition period, a set Q _i of treatment start dates is defined below.

Q _i = {t | TIME (t) = i, t∈ {1, ..., T}}
Next, the frame for one person starting on each start date t is calculated by the following procedures [1] to [3].

[1] A reference treatment process is set on the start date t.

[2] The start date of the remaining treatment process is calculated from the number of days left between the reference treatment process.

[3] From the start date of the irradiation treatment process, the number of irradiation treatment steps (determined by protocol j) is set for each day.

  The setting of the irradiation treatment process is performed in accordance with an irradiation schedule assignment method determined by each facility. For example, the following allocation method can be used. If the number of days that can be irradiated in that week is greater than the upper limit of the number of irradiations per week and the remaining number of irradiations is greater than or equal to the upper limit of the number of irradiations per week, the irradiation treatment process is suspended. Set by the allocation method defined by the treatment date pattern k. In other cases, the irradiation treatment process is sequentially assigned from the start date of the irradiation treatment process to the days after the next day when irradiation is possible.

For each start date of t∈Q _i , a frame for one person is calculated, and the number of persons for whom the irradiation treatment process is set on the day during each planning period is calculated. c ′ _tsikjl can be calculated by taking the work time of irradiation of the protocol j using the value obtained by calculating the work time of each day for the fixture creation and CT simulation when calculating the cost (FIG. 23). The work time of a treatment process in which work such as irradiation takes a plurality of days may be set by changing the work time depending on the number of times, and the multidimensional cost per person can be calculated in the same manner as the above calculation method.

  As described above, according to the present embodiment, for each allocation arc α of the allocation network, a flow rate limit value for limiting the flow rate as the number of assignable patients is calculated, and each allocation time type individually indicating the day type w Calculates the number of branches used for each allocation time type arc β that individually indicates the effective allocation time type node of the nodes “Month”,..., “Fri” for each protocol j, and allocates network, flow rate limit value, An allocation time type node “Month”,..., “Fri”, each allocation time type arc β and the number of branches used limit value are integrated to create a limit network, based on the limit network, treatment process information and constraint information With the configuration for generating the schedule pattern, it is possible to generate a regular and non-biased schedule pattern.

  Supplementally, in the present embodiment, the allocation time type nodes “Month”,..., “Fri” and the allocation time type arc β are generated, and each allocation time type arc β is limited to the number of branches to be used. In generating a weekly schedule pattern, it has an effect that weekly regularity can be easily created. The previous week and the next week are not necessarily the same due to the constraints of the original schedule, but if it is possible to generate a schedule that matches each week according to the day of the week, it is easy for the worker to learn the schedule pattern, and the burden on the worker is reduced. Decrease and increase safety. In addition, a rough plan in units of a plurality of months may be created by the above prior art, and a detailed plan for one day or a unit time unit may be created by the repetitive schedule generation device of this embodiment.

  In addition, even when work does not necessarily enter every week, in addition to creating a schedule with regularity that is easy for workers to recognize, regular weight adjustment of the objective function by the user is not necessary. Can be easily adjusted (FIGS. 24 and 25).

  For example, when there is no flow rate upper limit value FLOW (j) and use branch number upper limit value USE (j), as shown in FIG. The next week and the next week are assigned to different days of the week, and there is no regularity of the days of the week. Similarly, there is no regularity for other identical protocols “2” or “7”.

  On the other hand, when there is a flow rate upper limit value FLOW (j) and a use branch number upper limit value USE (j), as shown in FIG. And the next week are assigned to the same day of the week, and there is regularity of the day of the week. Other regular protocols “2” or “7” have the same regularity.

  Also, for example, when there is no flow rate upper limit value FLOW (j) and use branch number upper limit value USE (j), and the number of assigned persons varies depending on the day, the schedule pattern is the same protocol “ There is no regularity of the day of the week for the standard treatment process of “0”, “4” (Tue 2−Thu 2) or “7”. However, other same protocol “4” (Month 1-Month 2, Friday 1-Friday 2) has the regularity of the day of the week.

  On the other hand, when there are a flow rate upper limit value FLOW (j) and a use branch number upper limit value USE (j), and the number of assigned persons varies depending on the day, the schedule pattern is the same protocol “ There is regularity of the day of the week for the standard treatment process of 0 ”. Other regular protocols “4” or “7” have the same regularity.

  In the present embodiment, the case where the restricted network shown in FIG. 15 is formulated as a mixed integer linear programming problem based on the allocation network having the k and l options shown in FIG. 13 has been described as an example. However, the present invention is not limited to this, and the present invention can be implemented even when the restriction network shown in FIG. 14 is formulated as a mixed integer linear programming problem based on the assignment network having no k and l options shown in FIG. In this case, the holiday treatment day pattern k and the number of days of type 1 may be determined in advance. The same applies to the following embodiments.

<Second Embodiment>
FIG. 26 is a block diagram showing the configuration of the iterative schedule generation device according to the second embodiment. The same parts as those in FIG. 1 are denoted by the same reference numerals, and detailed description thereof is omitted. In the following. In addition, description of the overlapping part is abbreviate | omitted similarly also in each following embodiment.

  The second embodiment is a modification of the first embodiment, and is capable of adjusting the flow rate limit value and the use branch number limit value. The limit value setting unit is compared with the configuration shown in FIG. 13, a schedule pattern display unit 14 and a limit value display unit 15 are added.

  Here, the limit value setting unit 13 is a limit value setting function for individually setting a flow rate limit value and a use branch number limit value according to a user operation after the schedule pattern is output by the schedule pattern storage and output unit 12. A re-execution control function that controls the re-execution of the restriction network creation unit 10, the schedule pattern generation unit 11, and the schedule pattern storage and output unit 12 based on the set flow rate restriction value and use branch number restriction value I have.

Further, the limit value setting function of the limit value setting unit 13 represents the flow rate lower limit value FLOW_L (j) as shown in the following equation (4) when the flow rate lower limit value of the flow rate limit value is expressed as FLOW_L (j). It may have a flow rate lower limit setting function that calculates the flow rate limit value and sets the calculated flow rate lower limit value FLOW_L (j). It should be noted that the following formula (4) is calculated based on the flow rate upper limit value FLOW (j) when “FLOW (j) -1> 0” when calculating the flow rate lower limit value FLOW_L (j). -1 "and a procedure for obtaining" 0 "when" FLOW (j) -1≤0 "is included.

Similarly, the limit value setting function of the limit value setting unit 13 uses the number of used branches as shown in the following formula (5) when the used branch number lower limit value of the used branch number limit values is expressed as USE_L (j). A lower limit value USE_L (j) may be calculated as a used branch number limit value, and a used branch number lower limit value setting function may be provided to set the calculated used branch number lower limit value USE_L (j). It should be noted that the following equation (5) is obtained when “USE (j) −1> 0” based on the used branch number upper limit value USE (j) when calculating the used branch lower limit value USE_L (j). It also includes a procedure for determining as “0” when “USE (j) −1” and “USE (j) −1 ≦ 0”.

  The schedule pattern display unit 14 displays the current schedule pattern, and specifically has a function of displaying the schedule pattern output by the schedule pattern storage and output unit 12.

  The limit value display unit 15 displays the current flow rate limit value and the use number limit value, and specifically, the flow rate limit value calculated or set by the flow rate limit value calculation unit 7 and the number of used branches. The value calculation unit 9 has a function of displaying the used branch number limit value calculated or set.

  Next, the operation of the repetitive schedule generation device configured as described above will be described.

  Now, it is assumed that the operation described in the first embodiment is finished and a schedule pattern is output from the schedule pattern storage and output unit 12.

  At this time, the limit value display unit 15, as shown in FIG. 27, as the current limit value, the flow rate upper limit value FLOW (j), the flow rate lower limit value FLOW_L (j), and the used branch number upper limit value USE (j) The used branch number lower limit value USE_L (j) is displayed for each protocol j.

  Further, as shown in FIG. 28, the schedule pattern display unit 14 displays the output schedule pattern as the current schedule pattern.

  Here, when the user wants to proceed with leveling by loosening the regularity constraint, the limit value setting unit 13, as shown in an example in FIG. For example, the restriction is set by relaxing the use branch number upper limit value USE (j) of the protocol j “front 16” from 3 to 4.

  After that, the limit value setting unit 13 sends the use branch number upper limit value USE (j) “4” to the limit network creation unit 10 via the use branch number limit value calculation unit 9, thereby limiting the network creation unit. 10. Control is performed so that the schedule pattern generation unit 11 and the schedule pattern storage and output unit 12 are re-executed.

  As a result of the re-execution, the iterative schedule generation device has a better leveling than the schedule pattern shown in FIG. 28 because the variation of the schedule pattern increases due to the setting of the use branch number upper limit value USE (j) being loosened. That is, if a solution with a short maximum work time exists, a schedule pattern having a solution with a maximum work time of 4 hours is generated as shown in FIG.

  Conversely, when it is desired to increase the regularity by tightening the restrictions, the limit value setting unit 13 limits the flow rate lower limit value FLOW_L (j) in addition to the flow rate upper limit value FLOW (j) according to the user's operation. In addition to setting the value calculation unit 7, the use branch number lower limit value USE_L (j) is set in the use branch number limit value calculation unit 9 in addition to the use branch number upper limit value USE (j).

  The user applies the constraints of these lower limit values FLOW_L (j) and USE_L (j) only to a specific protocol j for which regularity is desired to be increased. For example, as shown in FIG. 31, when the constraints of the lower limit values FLOW_L (j) and USE_L (j) of protocol j “liver 8” are applied, a schedule pattern with increased regularity is obtained as shown in FIG. .

  The setting by the limit value setting unit 13 may be performed by the user manually inputting numbers, or by providing input screens such as a value increase button and a decrease button as shown in FIGS. The value may be automatically set when the value is increased or decreased.

  Further, the limit value setting unit 13 may set the flow rate lower limit value FLOW_L (j) calculated from the equation (4) or the like corresponding to the increased or decreased flow rate upper limit value FLOW (j) according to the user's operation. In the same way, even if the used branch number lower limit value USE_L (j) calculated from Equation (5) or the like is set corresponding to the increased or decreased used branch number upper limit value USE (j) according to the user's operation, Good.

  Similarly, the limit value setting unit 13 sets the flow rate lower limit value FLOW_L (j) and the used branch number lower limit value USE_L (j) to the flow rate limit value calculating unit 7 and the used branch number limit value calculating unit 9, respectively. The restriction network creation unit 10, the schedule pattern generation unit 11, and the schedule pattern storage and output unit 12 are controlled to be re-executed.

  As a result, a schedule pattern that includes restrictions on the flow rate lower limit value FLOW_L (j) and the usage branch number lower limit value USE_L (j) corresponding to the increase / decrease of the flow rate upper limit value FLOW (j) and the usage branch number upper limit value USE (j) Can be output.

  As described above, according to the present embodiment, after the output of the schedule pattern, the flow rate limit value and the use branch number limit value are individually set according to the user's operation, and the set flow rate limit value and the use branch number are set. In addition to the effect of the first embodiment, the flow rate limit value is controlled by the configuration in which the limit network creation unit 10, the schedule pattern generation unit 11, and the schedule pattern storage and output unit 12 are controlled based on the limit value. In addition, the regularity and the degree of bias in the schedule pattern can be adjusted by adjusting the restriction by the use branch number limit value.

  Supplementally, in the present embodiment, the allocation time type nodes “Month”,..., “Fri” and the allocation time type arc β are generated, and the schedule pattern is obtained by limiting the number of branches to be used for each allocation time type arc β. By calculating the upper limit value of the pattern type to be obtained, or the upper limit value and the lower limit value and applying constraints, it is possible to guarantee how much regularity is maintained on a weekly basis, and keep the number of patients as constant as possible ( The objective function of reducing the bias) can be optimized within a range that satisfies the constraints. In addition, there is an effect that the user can operate more easily than adjusting the weight of a plurality of objective functions. Furthermore, the setting of the flow rate limit value and the number of branches used limit value can be individually adjusted by the user in addition to the default settings shown in Equation (1), Equation (2), Equation (4), and Equation (5). This is possible, so it can be used when you want to loosen some restrictions.

<Third Embodiment>
Next, a repetitive schedule generation device according to the third embodiment will be described.

  The third embodiment is a modification of the first or second embodiment, and the above-described repetitive schedule generation device is replaced with radiation therapy, and a schedule for making a work easy to understand is prepared by a schedule of periodic work. It is the form applied when planning.

  Accordingly, the following points 1) to 14) are changed in the above-described repetitive schedule generation device.

  1) Use regular work instead of radiation therapy.

  2) Use each work type instead of each protocol.

  3) Instead of each treatment process, each work process is used.

  4) Use the number of work objects instead of the number of patients.

  5) Instead of the treatment site and the number of irradiation treatment steps, inspection work or tool replacement work is used.

  6) Instead of the restriction on the number of days between the treatment steps and the number of the irradiation treatment steps, the restriction on the number of days between the work steps is used.

  7) Use work points instead of day and week combinations.

  8) Use work point information instead of the allocation time information.

  9) Use workers instead of day types.

  10) The work point type information is used instead of the allocated time type information.

  11) A work point node is used instead of the allocated time node.

  12) A work type node is used instead of the protocol node.

  13) A work point type node is used instead of the allocated time type node.

  14) A work point type arc is used instead of the allocated time type arc.

  These changes of points 1) to 14) represent the following configurations.

  As shown in FIG. 33, the schedule pattern indicates each work point (allocation time) in the repetition period so as to indicate the date of the work process (allocation work) as a base point for each work type of the regular work. Work type is assigned. The schedule pattern includes each work type by the number of work objects of the work type. In FIG. 33, for ease of understanding, the type of worker is described at the top of the schedule pattern display, but this description may be omitted.

  The assigned work information storage unit (first storage unit) 2 stores assigned work information that defines each work type indicating inspection work or tool replacement work in each work process and the number of work objects of each work type. .

  The work point information storage unit (fourth storage unit) 3 stores work point information individually indicating work points to which work types can be assigned.

  The work point type information storage unit (fifth storage means) 4 stores work point type information that defines workers individually assigned to the work points in the work point information.

  The work process information is information that defines each work process. For example, information that defines each work process for each work type and a work process (hereinafter also referred to as a reference work process) as a base point of each work process. It is. In the standard work process, an upper limit of work time for performing the standard work process on each day can be specified. For other work processes, a schedule is generated at a relative position from the reference work process (the number of days left in between).

  The constraint information is information that defines constraints on work processes, and specifically, information that defines constraints on the number of days to be left between each work process.

  The allocation network creation unit 6 has a function of creating an allocation network based on the work point information and the allocation work information in the storage units 2 and 3.

  As shown in FIG. 34, the allocation network includes a plurality of work point nodes “1”,..., “12” that individually indicate work points indicated in the work point information, and each work specified in the assignment work information. A plurality of work type nodes “A”,..., “D” that individually indicate types, a plurality of assigned arcs α that individually indicate combinations of each work point node and each work type node, and work targets of each work type The numbers “6”,..., “1” are included.

  The flow rate limit value calculation unit 7 sets a flow rate limit value that limits the flow rate as the number of work targets that can be assigned for each assigned arc α based on the assigned work information and work point information in the storage units 3 and 5. It has a flow rate limit value calculation function.

Here, the flow rate limit value calculation function represents each work type specified in the assigned work information as j, represents the number of work objects of each work type j on the day during the repetition period as d _j, and is indicated in the work point information. A set of each work point to be expressed as I, the number of each work point included in the set I to be expressed as | I |, and a flow rate upper limit value among the flow rate limit values as FLOW (j), As shown in (1), a flow rate upper limit calculation function for calculating the flow rate upper limit value FLOW (j) as a flow rate limit value may be provided.

  The type element generation unit 8 is based on the work point information and the work point type information in each of the storage units 3 and 4, and a plurality of work point type nodes that individually indicate the workers and the valid among the work point type nodes. A function of generating a plurality of work point type arcs individually indicating various work point type nodes. The term “type element” may be read as “work point type node and work point type arc”.

  The used branch number limit value calculation unit 9 sets a used branch number limit value for limiting the number of used branch points of each work point type arc for each work type based on the allocation work information, the work point information, and the work point type information. It has a function to calculate the number of used branches to be calculated.

  Here, the use branch number limit value calculation function is obtained by dividing a value obtained by dividing the number of work points | I | indicated in the work point information by the number of workers | W | indicated in the work point type information h (= | I | / | W |) and when the used branch number upper limit value of the used branch number limit value is expressed as USE (j), as shown in the above-described equation (2), the used branch number upper limit value USE (j ) As a used branch number limit value may be provided.

  The restriction network creating unit 10 has a function of creating a restriction network by integrating the allocation network, the flow restriction value, each work point type node, each work point kind arc, and the use branch number restriction value. As shown in FIG. 35, the restricted network includes each work point node “1”,..., “12”, each work type node “A”,..., “D”, each assigned arc α, and the number of work objects. “6”,..., “1”, flow limit value (not shown), each work point type node (worker node in the figure) “a”,..., “C”, and each work point type arc β and the use branch number limit value (not shown) are included. The term “restricted network” may be read as “network having a work point type”.

  The schedule pattern generation unit 11 includes a schedule pattern generation function for generating a schedule pattern as shown in FIG. 33 based on the restriction network, work process information, and constraint information, and the generated schedule pattern as a schedule pattern storage and output unit. 12 has a schedule pattern writing function for writing to 12.

  The schedule pattern storage and output unit 12 has the functions described above.

  Next, the operation of the repetitive schedule generation device configured as described above will be described.

  For example, depending on the number of periodic tasks, there are a plurality of types of tasks that require periodic inspections and tool replacement, and the timing of replacement for each periodic task varies depending on the type of task.

  The work point is a timing at which a regular work may enter, which is set at a unit work interval that is a common divisor of the work intervals for each regular work. A schedule pattern can be generated by assigning work types to work points.

  In the second prior art, scheduling is performed in which regular work is always assigned at regular intervals. On the other hand, in this embodiment, as a schedule for repeating the periodic work for each repetition period of the work points, the periodic work does not necessarily have to be at regular intervals within the repetition period, and the work point at which the worker performs the work is the worker. If it is obtained for each, a schedule pattern that is easy to understand for workers can be generated.

  In performing such scheduling, the information input unit 1 assigns the assigned work information, the work point information, and the work point type information to the assigned work information storage unit 2, the work point information storage unit 3, and the work point according to user operations. Write to the type information storage unit 4.

  Similarly, the information input unit 1 writes work process information and constraint information into the schedule information storage unit 5 in accordance with a user operation.

  Subsequently, based on the work point information and the assigned work information in each of the storage units 2 and 3, for example, when the assignment network creation unit 6 has 12 work points and four work types in the repetition period, FIG. As shown in FIG. 4, a plurality of work point nodes “1”,..., “12” that individually indicate work points indicated in the work point information, and a plurality of work types that individually indicate each work type specified in the assigned work information. Work type nodes “A”,..., “D”, a plurality of assigned arcs α each indicating a combination of each work point node and each work type node, and the number of work objects “6”,. An allocation network including “1” is created.

  The multidimensional cost when assigning each work type to each work point can be calculated in the same way as in FIG. 23 if each work is repeated for each work point in the repetition period during the planning period. Leveling is possible.

  Here, as described above, the flow rate limit value calculation unit 7 calculates the flow rate limit value.

  Thereafter, the type element generation unit 8 integrates three work point type nodes (see FIG. 35) that are integrated into the restriction network as shown in FIG. 35 based on the work point information and the work point type information in the storage units 3 and 4. Middle, worker) “a”,..., “C” are generated, and arcs are generated between each work point type node and source node, and between each work point type node and work type node.

  In this way, when three work point type nodes “a”,..., “C” are generated, the use branch number limit value calculation unit 9 uses each work point type arc β as shown in FIG. The branch number upper limit value USE (j) can be calculated.

For example, in the case of the work type A, USE (j) is d _j / h (= 6) based on the equation (2) as d _j = 6, h = | I | / | W | = 12/3 = 4. / 4) Calculated as the smallest integer "2" or greater. In the case of FIG. 36, the number of work points that the worker a has three work points, the worker b five, and the worker c four is uneven, so that the number of repetitions is h = 4. Is not necessarily preferred. The method of calculating h described here is an example, and the present invention is not limited to this.

  Subsequently, the restriction network creation unit 10 integrates the allocation network, the flow restriction value, each work point type node “a”,..., “C”, each work point kind arc β, and the use branch number restriction value. A restricted network is created as shown in FIG.

  The schedule pattern generation unit 11 minimizes the evaluation value z of the above-described expression (3) based on the restriction network that can be formulated as described above, and the work process information and the constraint information in the schedule storage unit 5. By solving the mixed integer linear programming problem, a schedule pattern is generated as shown in FIG.

  Further, even when the number of workers and the assigned work points are changed as shown in FIG. 37, the upper limit value USE (j) of the number of branches used for each work point type arc β can be calculated as shown in FIG. As shown in FIG. 39, a schedule pattern can be generated.

  As described above, according to the present embodiment, compared to the first embodiment, the work point is used instead of the combination of the day of the week and the week, the work point information is used instead of the allocated time information, and the day of the week type is used instead. In addition to the effect of the first embodiment, by using a worker and replacing the allocated time type information with the work point type information, the worker who performs work at the work point is used as the work point type. It is possible to create a schedule pattern that facilitates gathering of specific work types at points.

  Further, three work point type arcs β corresponding to the three work point type nodes “a”,..., “C” are generated, and the number of used branches of these work point type arcs β is restricted, so that Since it is possible to easily create a schedule pattern in which the work amount is leveled and a specific work type easily gathers according to the work point type, it is possible to create a plan that is easier to understand for the worker.

  In addition, even when work points assigned to the same worker are given freely, a schedule can be created in which the same work is applied as much as possible to the same worker.

  The methods described in the above embodiments can be executed by a computer, and as programs that can be executed by the computer, magnetic disks (flexible disks, hard disks, etc.), optical disks (CD-ROM, DVD) Etc.), and can be stored and distributed in a storage medium such as a magneto-optical disk (MO) or a semiconductor memory.

  In addition, the device described in each embodiment can be implemented with either a hardware configuration or a combined configuration of hardware resources and software. As the software of the combined configuration, a program that is installed in a computer from a network or a storage medium in advance and causes the computer to realize the functions of the devices described in each embodiment is used. That is, the apparatus described in each embodiment may be implemented by executing the program in the memory of the computer by the processor of the computer.

  Moreover, although some embodiment of this invention was described, these embodiment is shown as an example and is not intending limiting the range of invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

  DESCRIPTION OF SYMBOLS 1 ... Information input part, 2 ... Allocation work information storage part, 3 ... Allocation time information storage part, 4 ... Allocation time type information storage part, 5 ... Schedule information storage part, 6 ... Allocation network creation part, 7 ... Flow limit value Calculation unit, 8 ... Type element generation unit, 9 ... Use branch number limit value calculation unit, 10 ... Limit network creation unit, 11 ... Schedule pattern generation unit, 12 ... Schedule pattern storage and output unit, 13 ... Limit value setting unit, 14 ... Schedule pattern display section, 15 ... Limit value display section, α ... Allocation arc, β ... Allocation time type arc.

Claims (12)

A repetitive type that generates a schedule pattern that assigns each protocol to indicate the base treatment process day for each radiotherapy protocol for each combination of day of the week and week in a repeating period that is repeated periodically A schedule generator,
A first storage means for storing each protocol indicating the treatment site and the number of irradiation treatment steps in each treatment step, and allocation work information defining the number of patients of each protocol;
A second storage means for storing treatment process information defining each treatment process for each protocol and a treatment process serving as a base point of each treatment process;
Third storage means for storing constraint information defining constraints on the number of days to be left between the treatment steps and the number of times of the irradiation treatment steps;
Fourth storage means for storing allocation time information individually indicating a combination of a day of the week and a week to which each protocol can be allocated;
Fifth storage means for storing assigned time type information defining day types individually assigned to combinations of days of the week and weeks in the assigned time information;
Based on the allocation time information and the allocation work information, a plurality of allocation time nodes individually indicating a combination of a day of the week and a week indicated in the allocation time information, and each protocol defined in the allocation work information are individually specified. An assignment network creating means for creating an assignment network including a plurality of protocol nodes to be shown, a plurality of assignment arcs individually indicating combinations of the respective assignment time nodes and the respective protocol nodes, and the number of patients of each protocol;
Based on the allocation work information and the allocation time information, a flow rate limit value calculating means for calculating a flow rate limit value for limiting the flow rate as the number of assignable patients for each of the allocated arcs;
Based on the allocated time information and the allocated time type information, a plurality of allocated time type nodes individually indicating the day type, and a plurality of individually indicating valid allocated time type nodes among the allocated time type nodes. Type element generating means for generating an allocated time type arc;
Based on the allocation work information, the allocation time information, and the allocation time type information, the number of used branches is limited to calculate a use branch number limit value that limits the number of branches used for each allocation time type arc for each protocol. A value calculating means;
By integrating the allocation network, the flow rate limit value, the allocation time type nodes, the allocation time type arcs, and the usage branch number limit values, the allocation time nodes, the protocol nodes, and the Restricted network creating means for creating a restricted network including an assigned arc, the number of patients, the flow rate limit value, each assigned time type node, each assigned time type arc, and the used branch number limit value ,
Schedule pattern generation means for generating the schedule pattern based on the restriction network, the treatment process information and the constraint information, and writing the generated schedule pattern in a schedule pattern storage means;
The schedule pattern storage means for storing the written schedule pattern;
A repetitive schedule generation device comprising: a schedule pattern output unit that reads a schedule pattern from the schedule pattern storage unit and outputs the read schedule pattern. The repetitive schedule generation device according to claim 1,
The flow rate limit value calculating means, said stands each protocol specified in assigned work information and j, the number of patients in each protocol j of the day during the repetition period represented as d _j, each indicated in the allocated time information When the set of days is represented by I, the number of each day included in the set I is represented by | I |, and the upper limit of the flow rate among the flow rate limit values is represented by FLOW (j), A flow rate upper limit calculating unit that calculates the flow rate upper limit value FLOW (j) as the flow rate limit value,
The used branch number limit value calculating means represents the number of types of weeks indicated in the allocated time information as h, and when the used branch number limit value of the used branch number limit value is expressed as USE (j), As shown in the following equation, an iterative schedule generation device comprising a used branch number upper limit value calculation unit that calculates the used branch number upper limit value USE (j) as the used branch number limit value.

The repetitive schedule generation device according to claim 2,
Limit value setting means for individually setting the flow rate limit value and the use branch number limit value according to a user operation after the output of the schedule pattern;
A re-execution control unit configured to control the re-execution of the restriction network creation unit, the schedule pattern generation unit, and the schedule pattern output unit based on the set flow rate restriction value and use branch number restriction value. A repeatable schedule generation device characterized by the above. The repetitive schedule generation device according to claim 3,
The limit value setting means includes:
When the flow rate lower limit value of the flow rate limit value is expressed as FLOW_L (j), as shown in the following formula, the flow rate lower limit value FLOW_L (j) is calculated as the flow rate limit value, and the calculated flow rate lower limit value FLOW_L a flow rate lower limit setting part for setting (j),
When the use branch number lower limit value of the use branch number limit value is expressed as USE_L (j), the use branch number lower limit value USE_L (j) is calculated as the use branch number limit value as shown in the following equation. And a use branch number lower limit value setting unit for setting the calculated use branch number lower limit value USE_L (j).

The repetitive schedule generation device according to claim 2,
Sixth storage means for storing planning period information indicating the planning period T;
A seventh storage means for storing day type information defining the day type l that is open between the treatment steps;
And 8th storage means for storing rest treatment day pattern information defining a pattern k of the day of resting from the radiation treatment step of each of the treatment steps,
The schedule pattern generation means represents a day in the planning period T as {0,..., T}, represents each treatment step as s, represents a set of the treatment steps s as S, and sets the set I as Each day included is represented by i, the set of patterns k is represented by K, the set of each protocol j is represented by J, the set of the number of days l is represented by L, and the day of the week is represented by w. , The set of the day of the week w is represented as W, the set of days of the day of the week w during the repetition period is represented as DAY (w), (w∈W), the day i, the pattern k, the protocol j, The number of patients in the combination ikjl with the day type l is expressed as y _ikjl, and the upper bound value of the value of the patient number y _ikjl is expressed as γ (where r = max {FLOW (j) | j∈J} ),

The upper bound value of the value that can be taken is represented by υ (where ν = max {USE (j) | j∈J}), and the working time of the treatment step s on day t when one patient is assigned in the combination ikjl C ′ _tsikjl is represented as a cost, _{1−1 variable} is represented as 1 when one patient is assigned in the combination ikjl, x _ikjl is represented as 1 and 0 when protocol j is assigned to the day of the week 0 −1 variable is expressed as u _wj , whole integer is expressed as Z, and evaluation value of the mixed integer linear programming problem shown in the following formula is expressed as z, the mixed integer linear programming problem so as to minimize the evaluation value z An iterative schedule generation device comprising solver means for generating the schedule pattern by solving

The repetitive schedule generation device according to claim 1,
Instead of the radiation treatment, using regular work,
Instead of each protocol, use each work type,
Instead of each treatment step, use each work step,
Instead of the number of patients, use the number of work objects,
Instead of the treatment site and the number of irradiation treatment steps, use an inspection operation or a tool replacement operation,
In place of the restriction on the number of days between the treatment steps and the number of irradiation treatment steps, use the restriction on the number of days between the work steps,
Instead of the combination of the day of the week and the week, using a work point,
Instead of the allocated time information, using work point information,
Instead of the day type, use workers,
Instead of the allocated time type information, using work point type information,
Instead of the allocated time node, use a work point node,
Instead of the protocol node, use a work type node,
Instead of the allocated time type nodes, using a working point type node,
An iterative schedule generating apparatus using a work point type arc instead of the allocated time type arc. Information input means, first storage means, second storage means, third storage means, fourth storage means, fifth storage means, allocation network creation means, flow rate limit value calculation means, type element generation means, Use branch number limit value calculation means, restriction network creation means, schedule pattern generation means, schedule pattern storage means, and schedule pattern output means, and radiation for each combination of day of the week and week in a repeating period that is repeated periodically It is a repetitive schedule generation method executed by a repetitive schedule generation device that generates a schedule pattern to which each protocol is assigned so as to indicate the date of a treatment process that is a base point for each protocol of treatment,
The information input means writes, in the first storage means, allocation work information that defines each protocol indicating the treatment site and the number of irradiation treatment steps in each treatment process, and the number of patients in each protocol. Process,
The information input means writes each treatment process for each protocol and treatment process information defining a treatment process as a base point of each treatment process in the second storage means;
The information input means writing into the third storage means constraint information that defines restrictions on the number of days and the number of times of the irradiation treatment process between the treatment steps;
The information input means writing into the fourth storage means allocation time information individually indicating a combination of a day of the week and a week to which each protocol can be assigned;
The information input means writing, in the fifth storage means, assigned time type information defining day types individually assigned to combinations of days of the week and weeks in the assigned time information;
Based on the allocation time information and the allocation work information, the allocation network creation means defines a plurality of allocation time nodes that individually indicate a combination of a day of the week and a week indicated in the allocation time information, and the allocation work information Create an assignment network that includes a plurality of protocol nodes that individually indicate each protocol that has been assigned, a plurality of assignment arcs that individually indicate a combination of each assignment time node and each protocol node, and the number of patients in each protocol The allocation network creation process to
A flow rate limit value calculating step in which the flow rate limit value calculating means calculates a flow rate limit value for limiting the flow rate as the number of assignable patients for each of the allocated arcs based on the allocation work information and the allocated time information. When,
The type element generation means includes a plurality of allocation time type nodes that individually indicate the day type based on the allocation time information and the allocation time type information, and an effective allocation time type among the allocation time type nodes. A type element generation step for generating a plurality of allocated time type arcs individually indicating nodes;
The usage branch number limit value calculating means limits the number of usage branches of each allocation time type arc for each protocol based on the allocation work information, the allocation time information, and the allocation time type information. A used branch number limit value calculating step for calculating a number limit value;
The limit network creation means integrates the allocation network, the flow rate limit value, the allocation time type nodes, the allocation time type arcs, and the use branch number limit value, and the allocation time nodes, A restriction network including each protocol node, each assigned arc, the number of patients, the flow rate limit value, each assigned time type node, each assigned time type arc, and the used branch number limit value A restriction network creation process to be created;
The schedule pattern generating means for generating the schedule pattern based on the restriction network, the treatment process information and the constraint information; and
A schedule pattern writing step in which the schedule pattern generation means writes the generated schedule pattern into the schedule pattern storage means;
The schedule pattern output means comprises a schedule pattern output step of reading a schedule pattern from the schedule pattern storage means and outputting the read schedule pattern. The repetitive schedule generation method according to claim 7,
In the flow rate limit value calculating step, each protocol defined in the allocation work information is represented as j, the number of patients of each protocol j on the day during the repetition period is represented as dj, and each day indicated in the allocation time information Is expressed as I, the number of each day included in the set I is expressed as | I |, and the flow rate upper limit value among the flow rate limit values is expressed as FLOW (j). A flow rate upper limit calculation step for calculating the flow rate upper limit value FLOW (j) as the flow rate limit value,
In the use branch number limit value calculating step, the number of types of weeks indicated in the allocated time information is represented as h, and the use branch number upper limit value of the use branch number limit values is represented as USE (j), As shown in the following equation, a recurring schedule generation method comprising a used branch number upper limit value calculating step of calculating the used branch number upper limit value USE (j) as the used branch number limit value.

The repetitive schedule generation method according to claim 8,
The repetitive schedule generation device further includes a limit value setting unit and a re-execution control unit,
A limit value setting step in which the limit value setting means individually sets the flow rate limit value and the use branch number limit value in accordance with a user operation after the output of the schedule pattern;
The re-execution control unit re-executes the restriction network creation step, the schedule pattern generation step, the schedule pattern writing step, and the schedule pattern output step based on the set flow rate restriction value and use branch number restriction value. And a re-execution control step for controlling the restriction network creation means, the schedule pattern generation means, and the schedule pattern output means so as to be executed. The repetitive schedule generation method according to claim 9,
The limit value setting step includes:
When the flow rate lower limit value of the flow rate limit value is expressed as FLOW_L (j), as shown in the following formula, the flow rate lower limit value FLOW_L (j) is calculated as the flow rate limit value, and the calculated flow rate lower limit value FLOW_L a flow rate lower limit setting step for setting (j);
When the use branch number lower limit value of the use branch number limit value is expressed as USE_L (j), the use branch number lower limit value USE_L (j) is calculated as the use branch number limit value as shown in the following equation. And a used branch number lower limit value setting step for setting the calculated used branch number lower limit value USE_L (j).

The repetitive schedule generation method according to claim 8,
The repetitive schedule generation device further includes sixth storage means, seventh storage means, and eighth storage means,
The information input means writing planning period information indicating the planning period T into the sixth storage means;
The information input means writing into the seventh storage means the day type information defining the day type l that is open between the treatment steps;
The information input means further comprises a step of writing into the eighth storage means the rest treatment day pattern information that defines the pattern k of the day when the radiation treatment step of each of the treatment steps is absent,
In the schedule pattern generation step, a day in the planning period T is represented as {0,..., T}, each treatment step is represented as s, a set of the treatment steps s is represented as S, and the set I Each day included is represented by i, the set of patterns k is represented by K, the set of each protocol j is represented by J, the set of the number of days l is represented by L, and the day of the week is represented by w. , The set of the day of the week w is represented as W, the set of days of the day of the week w during the repetition period is represented as DAY (w), (w∈W), the day i, the pattern k, the protocol j, The number of patients in the combination ikjl with the day type l is expressed as y _ikjl, and the upper bound value of the value of the patient number y _ikjl is expressed as γ (where r = max {FLOW (j) | j∈J} ),

The upper bound value of the value that can be taken is represented by υ (where ν = max {USE (j) | j∈J}), and the working time of the treatment step s on day t when one patient is assigned in the combination ikjl C ′ _tsikjl is represented as a cost, _{1−1 variable} is represented as 1 when one patient is assigned in the combination ikjl, x _ikjl is represented as 1 and 0 when protocol j is assigned to the day of the week 0 −1 variable is expressed as u _wj , whole integer is expressed as Z, and evaluation value of the mixed integer linear programming problem shown in the following formula is expressed as z, the mixed integer linear programming problem so as to minimize the evaluation value z A recurring schedule generation method comprising a solver step of generating the schedule pattern by solving

The repetitive schedule generation method according to claim 7,
Instead of the radiation treatment, using regular work,
Instead of each protocol, use each work type,
Instead of each treatment step, use each work step,
Instead of the number of patients, use the number of work objects,
Instead of the treatment site and the number of irradiation treatment steps, use an inspection operation or a tool replacement operation,
In place of the restriction on the number of days between the treatment steps and the number of irradiation treatment steps, use the restriction on the number of days between the work steps,
Instead of the combination of the day of the week and the week, using a work point,
Instead of the allocated time information, using work point information,
Instead of the day type, use workers,
Instead of the allocated time type information, using work point type information,
Instead of the allocated time node, use a work point node,
Instead of the protocol node, use a work type node,
Instead of the allocated time type nodes, using a working point type node,
An iterative schedule generation method using a work point type arc instead of the allocated time type arc.

Images