JP2014199552A - Delivery load leveling system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To derive a delivery schedule in which a work load relating to fuel delivery work is leveled, through data processing of a computer.SOLUTION: A delivery load leveling system comprises: databases DB2-DB5 for fuel container information, worker information, air temperature information and load information, respectively; and mathematical scheduling solution means 12. The mathematical scheduling solution means receives installation information of fuel containers, duty-related information of workers, air temperature information, transit time information and a work vehicle allowable delivery amount from the databases as input information, sets a remainder constraint condition, a duty constraint condition and a delivery amount constraint condition on the basis of the input information, and solves such delivery scheduling as to minimize a target function that sums up a penalty value indicating a degree of separation from a standard duty condition to which each of duty condition items for each worker in the duty constraint condition corresponds, as a mathematical scheduling question.

Description

  The present invention relates to a fuel supply mode in which a fuel container (cylinder) containing fuel such as LPG (liquefied petroleum gas) is installed at a customer's home to supply fuel, and replacement and recovery of the fuel container or fuel supply to the fuel container This is a fuel delivery work delivery plan in which a replacement fuel container or fuel for replenishment is mounted on a work vehicle and the customer's home is visited from the delivery base, and the work load of multiple workers The present invention relates to a delivery load leveling system that derives a delivery plan that is leveled over a predetermined planning target period by computer processing.

  As a system for accurately predicting the replacement or filling timing of a gas cylinder such as LPG and managing the replacement operation of the cylinder, there is one disclosed in Patent Document 1 or Patent Document 2 below.

  In the conventional system described in Patent Document 1, in order to improve the prediction accuracy of the replacement operation of the user gas cylinder and to be able to create a safe and rational cylinder delivery plan according to the prediction result, , Set the next gas cylinder replacement scheduled date from the user's past gas usage data, obtain the meter reading value of the gas meter for the scheduled user within the predetermined period, formulate the user group to be replaced with priorities, When selecting a delivery route for exchange, collect information such as delivery area traffic information, event information, and public facilities from the municipality website, etc., so that delivery can be performed in a short time. It has been proposed to set the route flexibly.

  In the conventional system described in Patent Document 2, a system for predicting an increase in gas usage that cannot be predicted only by past data on gas usage, and managing the replacement timing of the gas cylinder with high accuracy has been proposed. Yes. Specifically, the management system includes a measurement value input means for inputting a measurement value of a gas meter, an integrated usage amount deriving means for deriving an integrated gas supply amount after gas cylinder replacement based on the measurement value, a gas usage amount and temperature data Based on the association stored in the temperature usage relation database stored in the predetermined format, the temperature prediction input means for inputting the predicted temperature data from the external weather data providing device, and the predicted temperature data and the temperature usage relation database A gas supply prediction amount within a predetermined period is calculated and added to the integrated gas supply amount, and an integrated gas supply amount prediction means for calculating an integrated gas supply prediction amount within the predetermined period; and the integrated gas supply prediction amount is predetermined. An alarm generating means for generating a predetermined alarm when the threshold is exceeded is configured.

JP 2002-279025 A JP 2005-265739 A

  Delivery work for replacement and collection of gas cylinders such as LPG is highly seasonal, and due to the busy season in winter, it is possible to secure workers by irregular work, etc. Depending on the workload Therefore, there is a possibility that the bias of the work load on the worker cannot be sufficiently absorbed only by adjusting the necessary number of workers each time.

  In addition, by predicting gas consumption and making adjustments such as moving gas cylinder replacement ahead of schedule, leveling the workload during busy periods can be achieved by reducing overload during peak hours. In order to contribute to the reduction of labor costs, for the LPG operator, in order to improve profits, on the other hand, for the worker, to avoid overloading a specific worker, It was.

  By the way, in the conventional systems including those disclosed in the above-mentioned patent documents, there are many disclosures regarding a method for accurately predicting the replacement timing of the gas cylinder, creation of a retransmission route that can execute the formulated delivery work in a short time, and the like. However, no system has yet been proposed that can create a delivery plan for leveling the workload associated with delivery work.

  The present invention has been made in view of the problems related to the delivery work for replacement and recovery of the gas cylinders such as the above LPG, and the purpose thereof is to provide a delivery plan in which the work load related to the delivery work is leveled. It is to provide a distribution load leveling system derived by data processing.

In order to achieve the above object, the present invention provides:
Replacing and collecting fuel containers individually installed at a plurality of customer homes or performing customer home work such as refueling the fuel containers, mounting the fuel containers for replacement or fuel for replenishment on work vehicles, A delivery plan for a fuel delivery work performed by patroling the customer premises from a delivery base, wherein a delivery plan in which the work load of a plurality of workers is leveled over a predetermined target period is derived by computer processing A load leveling system,
A fuel container information database, an operator information database, an air temperature information database, and a road information database are formed in a storage area constituted by one or a plurality of storage devices, and mathematical operations are performed in the arithmetic processing unit constituting the computer. Planning solution means are configured,
The fuel container information database includes location information of the customer's home and installation information of the fuel container including attribute information of the installed fuel container, a delivery work date when the fuel delivery work is performed on the fuel container, and The fuel supply information of the fuel container including the remaining amount of the fuel container before the fuel delivery work is stored so as to be searchable for each customer.
The worker information database stores work-related information including standard work days of at least a month and standard work hours of one day for the worker in the planning target period so as to be searchable for each worker.
The temperature information database stores temperature information within at least the planning target period so as to be searchable by region of the installation location of the fuel container,
The road information database is the transit time information required to pass through each road within the work area including the delivery base and all the customer houses, or road information necessary to generate the transit time information. Store by date and road and searchable by road,
The mathematical program solving means is
The installation information and the fuel supply information read from the fuel container information database, the work related information read from the worker information database, the temperature information read from the temperature information database, read from the road information database, or Accepting as input information the transit time information generated based on the road information read from the road information database, and the deliverable amount that the work vehicle can deliver in one fuel delivery operation;
Based on the input information, the remaining amount restriction condition for the remaining amount of the fuel container in all the customer's homes not to disappear, the working condition items relating to working days and working hours for each worker during the planning period Work constraint conditions for satisfying the standard work conditions included in the work-related information, and the fuel delivery amount for all of the fuel containers allocated to one fuel delivery operation is less than the deliverable amount Set the delivery amount constraint condition as
Based on each constraint condition, with the remaining amount constraint condition and the delivery amount constraint condition as absolute conditions, the work condition item for each worker in the work constraint condition deviates from the corresponding standard work condition. The fuel container to be the target of the fuel delivery work for each worker on each work day during the planning period and the work order so that the objective function for summing up the first penalty value indicating the degree of A delivery load leveling system is provided which is configured to solve the delivery plan including a mathematical planning problem.

  In the delivery load leveling system of the above feature, the mathematical plan solving unit includes a fuel usage amount prediction unit, and the fuel usage amount prediction unit obtains the installation information and the fuel supply information from the fuel container information database by using the temperature. The temperature information is read from the information database, respectively, and based on the installation information, the fuel supply information, and the temperature information, a date that is provisionally set for each of the fuel containers during the solving process of the mathematical programming problem It is preferable to predict the remaining amount of the fuel container.

  In the delivery load leveling system of the above feature, the mathematical plan solving means is generated by the customer within the planning target period and within a predetermined grace period after the planning target period by using the fuel consumption prediction means. One or a plurality of limit days, which are days when the remaining amount of the fuel container runs out, is derived, and the fuel delivery work set during the solving process of the mathematical programming problem is performed for each customer. It is preferable that the set date is set within a recommended delivery period within the grace period from the limit date before the limit date corresponding to the delivery work setting date in the same order.

  In the delivery load leveling system of the above feature, the mathematical plan solving means has the delivery work setting date within a second grace period shorter than the grace days from the limit date before the limit date for each customer. Instead of an objective function that calculates a second penalty value indicating the degree of deviation from the second recommended delivery period on the delivery work setting date when it is not set within the second recommended delivery period, and aggregates the first penalty value Thus, it is preferable to derive the delivery plan so that an objective function for summing up the first penalty value and the second penalty value is minimized.

  In the delivery load leveling system having the above characteristics, it is preferable that an upper limit value is set for at least one of the first penalty value and the second penalty value individually or for the total.

  In the delivery load leveling system having the above characteristics, it is preferable that the second penalty value increases more greatly than a value proportional to the degree of the deviation as the deviation from the second recommended delivery period increases.

  In the delivery load leveling system having the above characteristics, it is preferable that the first penalty value increases more greatly than a value proportional to the degree of the deviation as the deviation from the standard work conditions increases.

  The delivery load leveling system of the above feature further includes a travel time database formed in the storage area, and the one customer's home and the other customer existing within a predetermined travel distance from the one customer's home. The shortest travel time along the movable route between the customer's home and the delivery base is a date / time within the planning target period based on the transit time information or the road information stored in the road information database. Separately, it is calculated in advance for all the customers and stored in the travel time database, and the mathematical plan solving means uses the shortest travel time instead of the passing time information as the input information. Is preferred.

  The delivery load leveling system according to the above feature is characterized in that the mathematical plan solving means takes the time required for one fuel delivery operation to deliver the installation location of the fuel container set in the one fuel delivery operation. It is preferable to calculate by totaling at least the total travel time from the base to the delivery base after traveling in the set order and the total work time of the customer home work at each of the fuel container installation locations. .

  In the delivery load leveling system having the above characteristics, it is preferable that the work condition items include the number of working days in a month, the working hours of one day, the number of continuous working days in the same month, and the number of continuous working days across the month.

  According to the delivery load leveling system having the above characteristics, the mathematical plan solving means uses the remaining amount constraint condition and the delivery amount constraint condition as absolute conditions based on the input information, and sets the work condition item for each worker in the work constraint condition. Since the delivery plan is derived so that the total of the first penalty values indicating the degree of deviation from the standard working conditions to which each corresponds is minimized, each of the working condition items for each worker is determined from the standard working conditions to which each worker corresponds. The delivery plan does not greatly deviate, and as a result, the workload based on the standard working conditions of each worker is leveled.

The block diagram which shows schematic structure of the delivery load leveling system which concerns on this invention LPG supply form from each LPG cylinder to each gas consuming device in each customer house related to the delivery plan targeted for the delivery load leveling system according to the present invention, and gas consumption amount in each gas consuming device in each customer house Of data aggregation form The flowchart which shows the derivation | leading-out procedure of the outline of the delivery plan of the fuel delivery operation | work by the delivery load leveling system which concerns on this invention The figure which shows an example of the main variables and constants used by the mathematical programming problem solved by the delivery load leveling system which concerns on this invention

  An embodiment of a delivery load leveling system according to the present invention (hereinafter referred to as “the present system” as appropriate) will be described with reference to the drawings. In the following description, the system of the present invention replaces and collects LPG cylinders (corresponding to fuel containers) individually installed at a plurality of customer homes (corresponding to customer home work), and fills the work vehicle with LPG. Equipped with a replacement LPG cylinder, a delivery plan for fuel delivery work that goes around the customer's home where the LPG cylinder is scheduled to be replaced on the day of the work from the delivery base, reduces the work load of multiple workers A mathematical programming problem that is leveled over a predetermined planning target period is solved and derived by computer processing.

  The delivery plan is configured as a set of individual delivery plans for each worker and work day. In each individual delivery plan, a worker, a work day, a customer home to be exchanged for one or more LPG cylinders, and a tour of the customer home The order is specified. The plan target period is a period in which an individual delivery plan is derived on each business day within the period, and LPG cylinder replacement occurs at least once at all customer homes and at least twice at some customer homes. Assuming the above-described period length, specifically, several months (for example, about 2 to 6 months) are assumed.

  As shown in FIG. 1, the system 1 of the present invention includes a fuel usage amount predicting means 11 and a mathematical plan solving means 12 in an arithmetic processing unit 2 constituting the computer, and is constituted by one or a plurality of storage devices 3. In the storage area, customer information database DB1, cylinder information database DB2 (corresponding to fuel container information database), worker information database DB3, temperature information database DB4, road information database DB5, travel time database DB6, and fuel consumption database It is configured with DB7. The computer system constituting the system 1 of the present invention may be constituted by dedicated computer hardware or a general-purpose computer system such as a personal computer.

  Each of the databases DB <b> 1 to DB <b> 7 is configured in an individual storage device 3, or several databases are combined into a single storage device 3. The database structure of each of the databases DB1 to DB7 is assumed to be a well-known structure, and operations such as registration, search, and deletion of data stored in each database are connected to the arithmetic processing device 2 or the storage device 3. This is performed by executing a database operation program installed in the other arithmetic processing unit. The retrieval of data necessary for the fuel usage amount predicting means 11 and the mathematical plan solving means 12 is respectively It is assumed that it is performed directly by the means 11 and 12 or via the operation program.

  The customer information database DB1 stores basic customer information such as a customer ID and customer attribute information (name, address, telephone number, etc.) in a searchable manner for each customer.

  The cylinder information database DB2 performs installation of LPG cylinders including location information (latitude and longitude, etc.) of customer premises and attribute information (type, number of installed units, etc.) of installed LPG cylinders, and fuel delivery work to the LPG cylinders. The fuel supply information of the LPG cylinder including the delivery work date, the number of LPG cylinders that performed the work, and the remaining amount of the LPG cylinder before the work is stored in a searchable manner for each customer.

  The worker information database DB3 stores work-related information of workers engaged in fuel delivery work during the planning target period so as to be searchable by worker. The work-related information includes at least the standard number of working days for a worker, the standard working hours for one day, a desired working day or a holiday. In this embodiment, an irregular work is assumed as a working form of the worker, and the number of working days per month and holidays, the working hours of one day, and the like are different for each worker and every month. In addition, the worker information database DB3 stores work results (working days, working hours of each working day, etc.) of each worker so as to be searchable by worker.

  The temperature information database DB4 stores at least temperature information within the planning target period so as to be searchable by region and day of the installation location of the fuel container. The temperature information assumes the use of daily temperature prediction data such as daily minimum temperature, daily maximum temperature, daily average temperature, etc., and uses temperature prediction data obtained from an external weather data provider as it is or after processing . If the obtained temperature forecast data is long-term forecast data for more than a month and is not daily temperature data, use data processed into daily temperature forecast data based on past daily temperature data such as normal temperature and the long-term forecast data. To do. Further, in the present embodiment, the daily temperature prediction data within a predetermined grace period after the planning target period and the first day of the planning target period are two days after the processing day for deriving the delivery plan (hereinafter simply referred to as “processing day”). In subsequent cases, daily temperature prediction data from the processing date to the planning target period is stored in the temperature information database DB4 for each region. Furthermore, at least the past one year of actual temperature data such as the daily minimum temperature, the daily maximum temperature, and the daily average temperature is stored in the temperature information database DB4 for each region.

  The road information database DB5 includes the transit time information required to pass through each road within the work range including the delivery base and all customer homes, and the road information necessary to generate the transit time information by date and time. It is stored so that it can be searched by road. In this embodiment, as a part of the road information, the past traffic conditions such as date / time and road, and road traffic information such as road construction and road closure sections and time zones are obtained from an external road traffic information provider. Obtained and stored in the road information database DB5. In addition, as other road information, road network information (intersections, distances between intersections, speed limits, etc.) within the above-mentioned work range, and locations of facilities such as schools that require caution when driving a vehicle or precautions regarding nearby roads Information is stored in the road information database DB5.

  In this embodiment, the transit time information of each road is, as basic data, a standard transit time obtained by dividing the distance between intersections by the standard operation speed set for the speed limit based on the road network information, It is calculated in advance and stored in the road information database DB5 by adding the passage time correction value for each date and time set in consideration of the road traffic information and the caution information. In addition, about 1 to 2 hours is assumed as a time unit according to the date and time. For roads that are scheduled to be closed and closed, and for roads that are scheduled to be closed, pass time information is not calculated and the road is connected to the road. Set.

  The travel time database DB6 is a date and time within a planning target period along a movable route between one customer's home and another customer's home and delivery base existing within a predetermined travel distance from the one customer's home. Another shortest travel time is stored. The shortest travel time is determined based on the customer home location information stored in the cylinder information database DB2 and the transit time information stored in the road information database DB5. (For example, calculated in units of 5 minutes) is used. In this embodiment, as an example, the daily delivery work time range (for example, 16 hours from 6 am to 10 pm) is divided into 8 time zones every 2 hours, and the shortest time by date. The travel time is stored for each day and time zone. Note that the time zone is a time zone for starting movement.

  The fuel consumption database DB7 includes actual consumption data of daily gas consumption (daily gas consumption) for each customer, day-of-day sensitivity data of gas consumption for each customer derived based on the actual consumption data, and Temperature sensitivity data is stored so that it can be searched by customer. As will be described later, the daily gas consumption is calculated based on the measured value of the gas meter 24 (see FIG. 2) installed in each customer's house. As day-of-week sensitivity data, for example, the daily fuel for each day of the week, national holidays, and specific days (family birthdays, etc.) specific to the customer relative to the monthly average of gas consumption on the first day of each month Assume a correction factor given by the ratio of consumption. As temperature sensitivity data, for example, the range of the daily minimum temperature (actual value) and the maximum value of the past actual value of gas consumption for the day when the daily minimum temperature was in the temperature range (for example, for 1 to 5 years) , Minimum value, average value, variance or standard deviation are stored in association with each other. The day sensitivity data and the temperature sensitivity data are updated, for example, periodically (for example, daily, weekly, monthly) or before the mathematical plan solving means 12 solves the mathematical plan problem of the delivery plan. And

  In this embodiment, the transit time information and the shortest travel time are separately derived before the mathematical plan solving unit 12 solves the mathematical plan problem of the delivery plan, and the mathematical plan solving unit 12 uses the derived shortest travel time. To do. In this embodiment, since the transit time information is used for calculating the shortest travel time, the transit time information may be stored in the travel time database DB6 without providing the road information database DB5. Further, the passing time information may be calculated only for roads and dates and times necessary for the calculation when calculating the shortest travel time.

  The fuel usage amount prediction means 11 reads the installation information and the fuel supply information for each customer from the cylinder information database DB2 by executing the fuel usage amount prediction program in the arithmetic processing unit 2, and sends each customer from the temperature information database DB3. On the other hand, the actual temperature data from the last day of the LPG cylinder replacement work (hereinafter referred to as the “most recent work day”) to the processing date, and the number of grace days from the processing date to the end date of the planned period The temperature information of the daily temperature prediction data until later is read out by region and by day, the actual consumption data, day sensitivity data and temperature sensitivity data are read out from the fuel consumption database DB7 by customer, and based on the read out data , The period from the last day of the most recent work day to the end of the planned period until the grace period (hereinafter referred to as “gas consumption calculation for convenience. It is configured to predict the day gas consumption of each customer in referred to as between "). The earliest day of the most recent work day is the earliest day of the last work day because the most recent work day is different for each customer. Note that the term “most recent work day” is used to clearly indicate the reference time point even when it means the date of the replacement work of the LPG cylinder performed immediately before that time point with respect to an arbitrary time point. . If there is no indication of the reference time point, the reference time point is the processing date.

  Although several methods can be considered as a method for predicting the daily gas consumption, in this embodiment, as an example, the gas consumption calculation period is divided into months included in the planning target period, and the planning target period is divided. The average daily gas consumption for each month is calculated using the gas consumption for the same month of the previous year as the monthly reference daily gas consumption. The daily gas consumption within the gas consumption calculation period is calculated on a daily basis by multiplying this correction coefficient by the temperature correction coefficient described later.

  For example, if the earliest day of the most recent work day is January 15, the processing date is January 30, the plan target period is February 1 to March 31, and the predetermined grace period is four days, The gas consumption calculation period is from January 15 to April 4, and the gas consumption calculation period is from January 15 to January 31, February 1 to February 28 (or 29th), It is divided into four months from March 1 to March 31, April 1 to April 4. The average value of the daily gas consumption from January 15 to January 31 of the previous year is used as the monthly reference day gas consumption from January 15 to January 31. Use the average daily gas consumption from February 1 to February 28 (or 29th) of the previous year as the monthly reference day gas consumption from February 1 to February 28 (or 29) . The same applies hereinafter.

  Next, an example of a method for calculating the temperature correction coefficient will be described. The actual value of the daily minimum temperature in the area where the customer's home that is the subject of calculation on the same day of the previous year corresponding to the gas consumption calculation period used for calculating the monthly standard day gas consumption is read from the temperature information database DB4, and each month (In the above example, the average daily minimum temperature in January to April is obtained, and the average value or average value and standard deviation of the actual gas consumption of the temperature sensitivity data corresponding to the average daily minimum temperature is calculated. The added value is calculated for each customer and month as a reference value for the previous month of daily gas consumption. Next, the actual measured value or predicted value of the daily minimum temperature of each day in the gas consumption calculation period is read, and the average value or average value and standard deviation of the actual value of gas consumption of the temperature sensitivity data corresponding to the daily minimum temperature are read. Is calculated as a reference value for the current day of daily gas consumption by customer and by day. By dividing the calculated reference value for the current day by the corresponding reference value for the previous month, the temperature correction coefficient for each customer and each day is calculated.

  The mathematical plan solving means 12 sets a plurality of constraint conditions, which will be described later, by executing a delivery plan derivation program including a general-purpose solution for mathematical programming problems in the arithmetic processing unit 2, and sets the plurality of constraint conditions. Below, it is configured to solve the individual delivery plan for each worker on each business day in the planning period as a mathematical planning problem so that the workload of multiple workers is leveled throughout the planning period. ing.

  FIG. 2 shows an LPG supply form from each LPG cylinder 21 to each gas consuming device 22 at each customer's house assumed in the present embodiment, and a data total form of gas consumption at each gas consuming device 22 at each customer's house. Is shown schematically. In FIGS. 2A and 2B, the LPG supply form is the same, but the data aggregation form is different.

  As shown in FIG. 2, the LPG filled in the LPG cylinder 21 at each customer's home is a regulator 23 that adjusts the gas supply pressure and the like, and a gas meter that measures the gas supply amount from the LPG cylinder 21 to the gas consuming device 22. 24, the gas is supplied to various gas consuming devices 22 owned by the customer (only one unit is simply shown in FIG. 2) by gas piping. Here, the gas meter 24 may be either a type that transmits a pulse signal every time a unit gas flow rate flows, or a type that measures a gas flow rate per unit time and transmits the measured value in a telegram format. As the gas flow rate measurement method, various measurement method gas meters applied to ordinary city gas gas meters such as a membrane method, a vortex flow rate measurement method, and an ultrasonic measurement method can be used.

  In the first data tabulation form, as shown in FIG. 2 (A), a data tabulation device 25 and a communication device 26 are installed at a customer's house, and the data tabulation device 25 is connected to a gas meter 24 via a signal line for communication. The device 26 is connected to the communication device 4 connected to the arithmetic processing device 2 via a communication line 27 (optical communication line, public telephone line, wireless data communication line, etc.). The data totalization device 25 receives the measurement value input from the gas meter 24 sequentially or periodically, calculates the daily gas consumption based on the measurement value, and periodically (for example, daily, weekly, or Every month), the data is transmitted to the arithmetic processing device 2 via the communication device 26, the communication line 27, and the communication device 4. The arithmetic processing device 2 stores the received daily gas consumption as the daily gas consumption for each customer in the fuel consumption database DB 7 for each customer and each day. The data aggregation device 25 may be configured by installing the aggregation processing software in a dedicated device specialized for the aggregation processing or a general-purpose computer device such as a personal computer installed in a customer's house. .

  In the second data tabulation form, as shown in FIG. 2B, a storage device 28 that stores the measurement value of the gas meter 24 is attached to the gas meter 24. Note that the storage device 28 may be attached either inside or outside the gas meter 24. The storage device 28 receives and stores the measurement value input from the gas meter 24 sequentially or periodically. The measured value stored in the storage device 28 is obtained by connecting the information processing terminal 29 and the storage device 28 carried by the worker with a predetermined communication cable when the operator replaces and collects the LPG cylinder 21. 29, and transmitted to the arithmetic processing unit 2 via the wireless data communication line 30. Alternatively, the measurement value captured by the information processing terminal 29 is calculated by connecting the information processing terminal 29 to the arithmetic processing device 2 via a data communication path such as a local network after the worker returns to the delivery base. You may make it transmit to the processing apparatus 2. FIG. In the arithmetic processing unit 2, the daily gas consumption for each customer is calculated based on the received measurement values, and stored in the fuel consumption database DB7 for each customer and each day.

  Further, the fuel supply information stored in the cylinder information database DB2 is, for example, every time the replacement work of the LPG cylinder is completed, the operator performs a predetermined input to the information processing terminal 29, and via the wireless data communication line 30, The data sent to the arithmetic processing unit 2 or the data taken into the information processing terminal 29 is arithmetically processed by the operator via the data communication path such as a local network after the worker returns to the delivery base. It connects to the apparatus 2 and transmits to the arithmetic processing apparatus 2. The arithmetic processing unit 2 stores the received fuel supply information in the cylinder information database DB2 for each customer.

  Next, a procedure for deriving a delivery plan for fuel delivery work by the system 1 of the present invention will be described using the flowchart of FIG. It is assumed that necessary data is stored in each of the databases DB1 to DB7 before starting the derivation process.

  First, the operator of the derivation process performs the input process of the first day (year / month / day) and the number of days (T) for setting the planning target period to the system 1 of the present invention via the input terminal of the computer ( Step # 1). The first day of the planning target period is set, for example, on the next day of the processing date.

  The mathematical plan solving means 12 sets a plan target period based on the input (step # 2), searches the cylinder information database DB2, derives the most recent work day, and sets a gas consumption calculation period. (Step # 3). In addition, 3 to 8 days (for example, 6 days) are used as a predetermined grace period from the last day of the planning target period to the end of the gas consumption calculation period. Here, if the first day of the planning period is set after the second day after the processing date, the number of days from the most recent work day to the first day of the planning period is longer than the minimum number of past delivery work intervals for the same customer. In such a case, since there is a possibility that the remaining amount of the LPG cylinder may run out before the planning target period in some customers, a warning message to that effect is generated on the display screen of the operator's input terminal. If the period from the most recent work day to the first day of the planning target period is included in the planning target period of the delivery plan derived from the previous process, the warning message is not necessarily displayed. It does not have to occur.

  Subsequently, the mathematical plan solution means 12 causes the fuel usage amount prediction means 11 to predict the daily gas consumption for each customer in the gas consumption calculation period (step # 4).

  Subsequently, the mathematical plan solving means 12 shows the following based on the data stored in each of the databases DB1 to DB7 and the daily gas consumption for each customer in the gas consumption calculation period predicted by the fuel usage prediction means 11. A constraint condition for the mathematical programming problem is set (step # 5). Then, a mathematical programming problem for minimizing an objective function described later under the constraint condition is executed by executing a program for solving the integer programming problem provided in the mathematical program solving means 12 to solve the mathematical programming problem. Then, a delivery plan for the fuel delivery work is derived (step # 6). By executing step # 6, the values of the variables used to formulate the constraint conditions and the objective function are determined, and each work day during the planning target period and each worker's individual delivery plan (which can be determined by one fuel delivery work, The order in which the customer's home is circulated is determined in a state where the work loads of a plurality of workers are leveled. In this embodiment, an integer variable is adopted as a variable used for formulation of constraint conditions and objective functions, and an existing integer programming problem solving program is used as a mathematical programming problem solving program. A detailed description of itself is omitted.

  Next, constraint conditions, objective functions, and variables and constants used for formulation of mathematical programming problems will be described.

  As variables and constants for defining the objective function and constraint conditions of the mathematical programming problem, for example, binary variables, discrete variables, and constants shown in FIG. 4 are used. In FIG. 4, only main variables and constants are listed. The subscripts attached to each variable and constant are defined as follows. i and j are ordinal numbers (natural numbers) for identifying customer premises. However, i, j = 0 indicates a delivery base. k is an ordinal number (natural number) for identifying the worker. In addition, t is an ordinal number (natural number) indicating what day of the planning target period (T days), s is an ordinal number (natural number) indicating a time zone dividing the delivery work time range of one day, and r Is an ordinal number (natural number) indicating the month in which the plan target period is divided by month, and q is an ordinal number (natural number) indicating the day of the gas consumption calculation period. Subscripts other than variables and constants will be explained each time.

α _{t, i} is a binary variable indicating whether or not the delivery work to the customer's home i is performed on the t-th day (hereinafter abbreviated as time t) of the planning target period. β _{t, k} is a binary variable indicating whether or not the worker k performs delivery work at time t. γ _{t, i, j, k} is a binary variable indicating whether or not the worker k moves from the customer home (or delivery base) i to the customer home (or delivery base) j at time t. , I ≠ j. δ _{t, i, j, k} is a discrete variable (variable value is 0) indicating the time zone when worker k starts moving from customer (or delivery base) i to customer (or delivery base) j at time t. ~ 8), i ≠ j. When there is no movement from the customer (or delivery base) i to the customer (or delivery base) j by the worker k at the time t, the variable value is zero. [epsilon] _{t, i} is a discrete variable indicating which day in the gas consumption calculation period the most recent work day based on the time point t at the customer's home i. The binary variable has a variable value of 1 if the content of each variable is true, and 0 if not. Moreover, the time point t of the variable β _{t, k} is not limited to the planned period, but covers the period including the entire period of the first month of the planned period. For the period before the processing date of the first month of the plan target period _, the variable value of β _{t, k} is set based on the work record of worker k stored in the worker information database DB3. The constants will be described in detail in the description of the constraint conditions below.

  Next, the constraint conditions will be described. In the present embodiment, four types of constraint conditions are assumed as the constraint conditions: a remaining capacity constraint condition, a work constraint condition, a delivery amount constraint condition, and a constraint condition between variables.

1. About remaining capacity restriction conditions:
The remaining amount constraint condition is that the remaining amount of the LPG cylinder does not become 0 throughout the planning target period at each customer house i. It is specified that the given remaining amount B _{t, i} does not become 0 or less (B _{t, i} > 0). Even if the remaining amount restriction condition is not satisfied at all time points t, it is sufficient if the remaining amount restriction condition is satisfied on the delivery work day of the customer home i where α _{t, i} = 1. B0 _{i in} the first term on the right side of Equation 1 is a constant indicating the gas capacity when the LPG cylinder installed in the customer's house i is full. U _{i, q in Equation} 1 is the daily gas consumption calculated by the fuel usage prediction means 11 at the customer's home i on the qth day of the gas consumption calculation period. T0 is a constant indicating the difference in days from the first day of the gas consumption calculation period to the first day of the planning period. The second term on the right side of Equation 1 is the customer's home from the most recent work day (ε _{t, i} day of the gas consumption calculation period) to the time t (t + T0 day of the gas consumption calculation period) with reference to the time t. The accumulated value of daily gas consumption in i is shown.

In the present embodiment, the second remaining amount constraint condition shown in the following formula 2 is further adopted. First, for each customer home i, the time horizon remaining C _t during a period from the first day until the gas consumption calculation period last _{day, i} becomes zero day (hereinafter, referred to as "critical date tx i _(m)") Are derived using the fuel usage amount predicting means 11. M in the parenthesis is a natural number and indicates the number of limit date counted from the first day of the planning target period. Then, the customer premises i Separately, the number of days difference between the limit date _tx i last working day relative to the (m) (epsilon _txi gas consumption calculation period _{(m), i-th} day) and limit date _tx i (m) is As shown in Equation 2 below, the second remaining amount constraint condition is that the number of days is C or less. Due to the second remaining amount restriction condition, the delivery work setting date (time t when α _{t, i} = 1) set to the customer's home i set during the solution processing corresponds to the delivery work setting date in the same order. It is set within the recommended delivery period within the number of grace days C from the limit date before the limit date tx _i (m). If the grace period C is set too short _, the possible range of the latest work day ε _{txi (m), i} is limited, and as a result, the degree of freedom of optimization described later is limited, and all constraints are satisfied. There is a possibility that an optimal solution cannot be obtained.

2. About work restrictions:
The work constraint condition is a standard work condition in which work condition items relating to work days and working hours for each worker k in the planning period are included in work-related information stored in the worker information database DB3 ( (Standard working days per month, standard working hours for one day, desired working days, holidays, etc.). As working condition items, for example, the working day of the first month of the rth month of the planning period, the number of working days and the continuous working days of the first month of the rth month of the planning period, continuous spanning from the end of the month to the next month Assume the number of working days, holidays, etc., within the planning target period that operator k wants to obtain. In the present embodiment, the work constraint condition is set for each work condition item described above and for each worker k.

The first work constraint condition is a work constraint condition related to the working hours on the first day of the r-th month for the worker k, and is given by the following equation (3). Here, P1 _{t, k} in Equation 3 indicates a penalty value in the first work constraint condition of the r-th month at time t for worker k, and P1UL is the penalty value P1 _{r, k} This is the upper limit. D _{t, k} in _Equation 3 represents the total number of working hours per day for worker k at time t, and is given by Equation 4. Further, E _{r, k} in Equation 3 is a constant representing the standard working time of the first day of the rth month, and is stored in the worker information database DB3. In addition, _{r of} the constant _{Er, k} indicates in which month of the planning period the month to which the time point t belongs. A0 _{i in} the first term on the right side of Equation 4 is a constant indicating the time required for replacement and collection of the LPG cylinder at the customer's home i, and A1 _{t, i, j, in} the second term on the right side of Equation 4 _δ is a constant (i ≠ j) indicating the time required to move from the customer (or delivery base) i to the customer (or delivery base) j in the time zone s (s = δ _{t, i, j, k} ) at time t. ) And equal to the shortest travel time stored in the travel time database DB6. D0 in the third term on the right side of Equation 4 is used for preparation and clean-up of delivery work other than the time required for exchanging and collecting LPG cylinders within the working hours of the day, and the travel time between customer premises and between customer premises and delivery base This is the total time required and break time.

The second work constraint condition is a work constraint condition related to the number of working days in the first month of the r-th month for the worker k, and is given by the following formula (5). Note that P2 _{r, k} in Equation 5 indicates a penalty value in the second work constraint condition of the r-th month for the worker k, and P2UL is an upper limit value of the penalty value P2 _{r, k.} . Moreover, beta _{t, k} addition operation Σ of t = t (r) means that adding all beta _{t, k} in the r Month time t. In the second work constraint condition shown in Equation 5, when r = 1, the entire period of the first month of the planning target period is included. In Formula 5, F1 _{r, k} is a constant indicating the standard working days of the first month of the r-th month for the worker k, and is stored in the worker information database DB3.

The third work constraint condition is a work constraint condition related to the continuous working days in the r-th month for the worker k, and is given by the following formula (6). Note that P3 _{r, k, n in} Equation 6 indicates the n-th penalty value in the third work constraint condition of the r-th month for the worker k, and S1 _{r, n} (β _{t, k} | F2 _{r, k} ) is the number of consecutive days when the time t in the r-th month when the discrete variable β _{t, k} becomes 1 exceeds F2 _{r, k,} and is the nth in the r-th month. It is a function that gives the number of consecutive days that occur, and P3UL is the upper limit value of the penalty value P3 _{r, k} . Further, F2 _{r, k} in the equation 6 is a constant indicating the standard continuous working days in the r-th month for the worker k, and is stored in the worker information database DB3. In the third work constraint condition of Equation 6, when r = 1, the entire period of the first month of the planning target period is included.

The fourth work constraint condition is a work constraint condition related to the number of continuous working days straddling from the end of the month to the next month for the worker k, and is given by the following equation (7). Note that P4 _{k, p} in Equation 7 indicates the p-th penalty value in the fourth work constraint condition for worker k, and S2 _p (β _{t, k} | F3 _k ) is a discrete variable β _{t, This} is a function that gives the number of consecutive days when the time t at which _k becomes 1 exceeds F3 _k and continues from the end of the month to the next month, and gives the p-th consecutive days that occur in the planning target period, and P4UL is This is the upper limit value of the penalty value P4 _{k, p} . Further, the F3 _k in Equation 7, a constant indicating the continuous working days straddling over the next month from the end of the standard for operator k, stored in the operator information database DB3. In the fourth work constraint condition of Equation 7, when r = 1, the entire period of the first month of the planning target period is included.

The fifth work constraint condition is a work constraint condition related to a holiday within the r-th month of the plan target period that the worker k desires to obtain, and is given by the following equation (8). In addition, P5 _{r, k} in Equation 8 indicates a penalty value in the fifth work constraint condition for the worker k, and P5UL is an upper limit value of the penalty value P5 _{r, k} . G _{t, k} in _Equation 8 is a binary constant indicating whether or not it is a holiday desired by the worker _k at the time point t. When 1, the worker k wants to acquire a holiday at the time point t. Indicates that The constants G _{t, k} are stored in the worker information database DB3. Moreover, beta _{t, k} and _{G t,} the addition operation of the product of _k sigma of t = t (r) is, beta _{t, k} and _{G t} in the r Month time _t, adding all the products of _k Means. In the fifth work constraint condition shown in Equation 8, when r = 1, the entire period of the first month of the planning target period is included. In the fifth work constraint condition, it is assumed that each worker has 0 days as a standard number of days on which a work day is a holiday desired to be acquired every month. Therefore, the total number of days in which the desired holiday of each month becomes the work day is the penalty value P5 _{r, k} in the fifth work constraint.

3. About delivery volume constraints:
The delivery amount constraint condition is a constraint condition in which the fuel delivery amount for all of the LPG cylinders of the customer's house assigned to one fuel delivery operation is equal to or less than a predetermined deliverable amount. In the present embodiment, the fuel delivery amount and the deliverable amount are defined as the number and total weight of replacement LPG cylinders that can be mounted on one work vehicle in one fuel delivery operation. Therefore, in this embodiment, the first delivery amount constraint condition related to the number of LPG cylinders and the second delivery amount constraint condition related to the weight of the LPG cylinder are used as the delivery amount constraint conditions. Here, it is assumed that each worker performs the fuel delivery work only once a day. In other words, it is assumed that the departure from the delivery base, each customer's home and return to the delivery base, and the one day fuel delivery work is completed, and the replacement LPG cylinder is refilled on the way back to the delivery base. Assume that you do not. Each worker uses a work vehicle having the same loading capacity.

The first delivery amount constraint condition is given by the following Equation 9 and the second delivery amount constraint condition is given by the following Equation 10 for one fuel delivery operation of the worker k at time t. Here, H1 _i in Equation 9 is a constant indicating the number of LPG cylinders installed at customer's house i, and H1UL is a constant indicating the upper limit of the number of LPG cylinders that can be mounted on one work vehicle. And is stored in the cylinder information database DB2. In addition, H2 _i in Formula 10 is a constant indicating the total weight of the LPG cylinder installed at the customer's home i, and H2UL indicates the upper limit value of the total weight of the LPG cylinder that can be mounted on one work vehicle. It is a constant and is stored in the cylinder information database DB2.

4). About constraints between variables:
In the present embodiment, three types of binary variables, α _{t, i} , β _{t, k} , and γ _{t, i, j, k} , and δ _{t, i} are defined as variables for defining each constraint condition. _{, J, k} and ε _{t, i} are defined as two types of discrete variables, and α _{t, i} and β _{t, k} are determined from γ _{t, i, j, k} in the planning period, Since it is derived uniquely from the following equations 11 and 12, it is not an independent variable.

Equation 11 shows that at any time t, any worker k moves from customer home i to any one other customer home j, and any other customer home j moves from customer home i. If it occurs, the movement will only occur once, that is, the variable γ _{t, i, j, k} may be 1 at any time t and any worker k. For example, it means that for a specific customer home i, there is one corresponding customer home j or delivery base (j = 0), and the value on the right side of each equal sign in Equation 11 is 1. The value of the variable α _{t, i} is 1. On the other hand, if the variable γ _{t, i, j, k} does not become 1 for a specific customer house i at the time t, naturally, the value of the variable α _{t, i} becomes 0.

Further, the operator Max (γ _{t, i, j, k} | i, j) on the right side of Expression 12 is a variable γ indicating whether or not the worker k moves between the customer premises or between the customer premises and the delivery base at the time t. The maximum value of a plurality of variable values defined by combinations of all _{i, j of t, i, j, k} is derived, and the variable β _{t, k} includes 1 in the plurality of variable values. 1 and 0 when all are 0.

Next, a constraint condition between variables γ _{t, i, j, and k} will be described. As shown in the first formula, the following equation 13 is used for variables γ _{tx, i, j, kx} indicating whether or not a certain worker k _x moves between customer premises or between a customer premises and a delivery base at a certain time t _x . , if the combination of the variable value is 1 the customer's home (or delivery base) _{i x} and _{j x} exists _{(γ tx, ix, jx,} kx = 1), at the same time _{t x,} the operator _{k x} That there is no reverse movement between the customer premises (or delivery bases) i _x and j _x (Equation 2), and at the same time t _x , the worker k _x including all workers k, the customer's home (or delivery base) customer premises from i _x (or delivery base) that move to the other j _x is not present (third type), at the same time t _x, the working For all workers k including operator k _x Kyakutaku (or delivery base) the customer's house from outside i _x (or delivery base) the movement of the j _x is not present (fourth type), and, at the same time t _x, the work other than the worker k _x This indicates that there is no movement from the customer home (or delivery base) i _x to the customer home (or delivery base) j _x in the person k (Formula 5).

Further, as shown in the first formula, the equation 14 represents a variable γ _{tx, i, j,} which indicates whether or not a certain worker k _x moves between customer premises or between a customer premises and a delivery base at a certain time t _{x .} If there is a combination of customer homes (or delivery bases) i _x and j _x whose variable value is 1 at _kx (γ _{tx, ix, jx, kx} = 1), the same time point t _x and the same worker k _{In x} , there is always one movement from the customer home (or delivery base) h other than the customer home (or delivery base) i _x and j _x to the customer home (or delivery base) i _x (the second formula) ), at the same time _{t x} and the same worker _{k x,} transfer from the customer's home (or delivery base) _{j x} to the customer's home (or delivery base) _{i x} and _{j x} other than the customer's home (or delivery base) g Must always exist (Equation 3), at the same time In _x and the same worker _{k x,} moves from the delivery base (i = 0) to the customer premises j (j ≠ 0) is present exactly one (fourth type), and, the same time _{t x} and the same work This shows that there is always one movement from the customer's home i (i ≠ 0) to the delivery base (j = 0) in the person k _x (Equation 5). As a result, as a moving route between customer premises or between customer premises and a delivery base by a worker k _x at a certain time t _{x, a} single route that returns from the delivery base to the delivery base via one or more customer premises. Although the travel route is determined, it is desirable to set as another constraint condition that there is no circulation route other than the travel route that does not depart from the delivery base, does not return to the delivery base, and circulates between customer premises.

Furthermore, the discrete variable δ _{t, i, j, k} indicating the time zone when the worker k starts moving from the customer (or delivery base) i to the customer (or delivery base) j at the time _t is given by Introduced for the purpose of defining work constraint conditions, it can be derived from binary variables γ _{t, i, j, k} , constant A0 _i, and constant A1 _{t, i, j, δ} . Specifically, the operator k and a given time t, gamma _{t, i, j,} extracts k = 1 to become customer home i, gamma _{t, i, j,} k = 1 and becomes i, From the relationship of j, the traveling order J _{t, i, k} (natural number) of the customer home i that is the target of the fuel delivery work performed by the worker k at the time t is determined, so the J _{t, i, k} th customer The time JT _{t, i, k at} which the movement from the home i to the J _{t, i, k} + 1st customer home j starts is J _{t, i, k} -1 from the first customer home h to J _{t, i, k} Since it is after (A0 _i + A1 _{t, h, i, δ} ) from the time JT _{t, h, k at} which the movement to the customer's home i starts, it is expressed by the recurrence formula shown in the following formula 15. . When J _{t, i, k} = 1, the time JT _{t, h, k} is the time when the worker k starts the fuel delivery operation at the time t. Therefore, by solving the recurrence formulas in order, each time JT _{t, i, k} is calculated, and a discrete variable δ _{t, i, j, k} indicating the time zone is calculated.

Further, the discrete variable ε _{t, i} indicating the number of days in the gas consumption calculation period corresponding to the most recent work day at the time point t at the customer home i is convenient for defining the second remaining amount constraint condition. It is derived from the binary variable α _{t, i} and the constant T0. Specifically, as shown in the following Expression 16, at an arbitrary time tx (tx <t) before the time t _, the maximum value of tx that satisfies α _{tx, i} = 1 (in Expression 16, the operator Max) The value obtained by adding T0 to the value of the discrete variable ε _{t, i} .

As described above, among the above five types of variables, the independent variable is only one type of binary variables γ _{t, i, j, k} . In fact, for a given time t and worker k, the individual delivery plan of worker k at time t is uniquely identified by the set of customer homes i, j where γ _{t, i, j, k} = 1. Therefore, the binary variable γ _{t, i, j, k} is a decision variable in the mathematical programming problem in this embodiment. Note that the number of the decision variables γ _{t, i, j, k} is the number of combinations of the number of days in the planning target period, the number of customer premises, and the number of workers. The number of will be enormous. Therefore, in such a case, the number of variables γ _{t, i, j, k} can be reduced by grouping the customer premises and assigning each worker to a part of the customer premises groups.

  Furthermore, the constraint condition between the variables defined by the relational expressions shown in Expressions 11 to 16 can be handled as a rule that defines a dependent variable that is provisionally used in other constraint conditions.

Next, the objective function will be described. In this embodiment, in order to level the work load of a plurality of workers, as an example, the five penalty values P1 _{t, k} , P2 _{r, k} , P3 _{r, of the} first to fifth work constraint conditions described above are used. _An objective function P given by aggregating quadratic polynomials _{k, n} , P4 _{k, p} , P5 _{r, k} is defined as shown in the following equations 17 to 22. Note that m1 to m5 and n1 to n5 in Equations 18 to 22 are weighting coefficients for the penalty values and are positive values.

The combination of binary variables γ _{t, i, j, k} that minimizes the value of the objective function P is determined by the mathematical programming problem solving process executed in steps # 5 and # 6 of the flowchart of FIG. . As described above, for a given time point t and worker k, the individual delivery plan of worker k at time point t is unique by the set of customer homes i, j with γ _{t, i, j, k} = 1. And a delivery plan for all workers in the entire planning target period is determined by the set of the individual delivery plans. Here, Q1 to Q5 (each corresponding to the first penalty value) of each term of Equation 17 is the respective corresponding first to fifth work constraint conditions as shown in Equation 18 to Equation 22. Since each coefficient is a positive number given by the quadratic value of the penalty value, it is corrected so that the larger the deviation from the standard working conditions, the larger the value proportional to the degree of the deviation. It is a penalty value. Therefore, in each sum of Q1 to Q5 of each term on the right side of Equation 17, even if the penalty values P1 to P5 are smaller on average, even if the sum of the primary terms is the same, the value of the secondary term is As the penalty value becomes smaller and the individual penalty values vary, the value of the second-order term becomes larger, and the corrected penalty values Q1 to Q5 become larger. As a result, since the objective function P has quadratic terms of the individual penalty values P1 to P5, the penalty values P1 to P5 are averaged for the worker k by minimizing the objective function P. It will be minimized and meet the purpose of leveling the workload of multiple workers.

Hereinafter, another embodiment will be described.
<1> In the above embodiment, the penalty values Q1 _{t, k} , Q2 _{r, k} , Q3 _{r, k, n} , Q4 _{k, p} based on the work constraint condition expressed by the above equation 17 are used as the objective function of the mathematical programming problem. , Q5 _{r, k} are used in total, but the objective function is not limited to that shown in Equation 17. For example, as a penalty value based on the work constraint condition, only a part may be used without using all of Q1 to Q5, and for some of Q1 to Q5, the corresponding penalty values P1 _{t, k} , Only the primary terms of P2r _{, k} , P3r _{, k, n} , P4k _{, p} , P5r _{, k} may be used. Furthermore, you may add the penalty value regarding work condition items other than the above.

Furthermore, in addition to various penalty values (first penalty value) based on the work constraint condition, a penalty value (second penalty value) related to the remaining amount constraint condition is defined, and the first and second penalty values are defined. It is good also as the objective function. In this case, it is also preferable to use the second penalty value Q6 _{i, m} given by, for example, the following equation (23). Note that P6 _{i, m} in Equation 23 is obtained by subtracting a constant C0 (second grace day) from the left side of Equation 2 indicating the second remaining amount constraint condition, as shown in Equation 24. M6 and n6 are weighting factors, respectively. The constant C0 in Equation 24 is a value (0 ≦ C0 <C) that is greater than or equal to 0 and smaller than the grace period C on the right side of Equation 2. Further, the objective function P when the second penalty value Q6 _{i, m} is used is given by Equation 25.

<2> In the above embodiment, the above-mentioned α _{t, i} , β _{t, k} , γ _{t, i, j, k} , δ _{t, Although i, j, k} , and ε _{t, i} are used, the variables to be used are not limited to the variables described above. Moreover, the formulation of the constraint condition and the objective function is appropriately changed depending on the variable to be used.

For example, use of a binary variable γ _{t, i, j, k} indicating whether or not the worker k moves from the customer home (or delivery base) i to the customer home (or delivery base) j at the time _t. Instead of this, a binary variable γ _{t, i, k} indicating whether or not the worker k performs delivery work to the customer's home i at time t may be used. In this case, when there are a plurality of customer homes i with γ _{t, i, k} = 1 for a specific time t and worker k, it is necessary to determine the order of the fuel delivery work for the plurality of customer homes i. However, the order can be set in one way depending on constraints such as constraints and minimization of the objective function. However, the constraint conditions and the formulation of the objective function are changed by using the variables γ _{t, i, k} . Further, instead of using the variable γ _{t, i, k} , for example, the variable J _{t, i} indicating the order in which the customer's home i is subjected to the fuel delivery work performed by the worker k at the time t used in Equation 15 is used. _{, K} may be used. In this case, J _{t, i, k} = 0 means the fuel delivery work to the customer's home i by the worker k at the time t.

  Furthermore, even when the variables used are the same as those in the above embodiment, the constraint condition and the objective function are not necessarily limited to those in the above embodiment.

<3> In the above embodiment, the fuel usage amount prediction means 11 calculates the daily gas consumption for each customer from the daily gas consumption (actual consumption data) in the fuel consumption database DB7, and the monthly reference daily gas consumption. As an example, a case is described in which the correction coefficient of the day sensitivity data calculated in advance and the temperature correction coefficient are read out from the fuel consumption database DB7 and multiplied by the monthly reference day gas consumption. The prediction of consumption is not limited to the method described above.

  For example, even if the accuracy of the daily gas consumption for each customer is low, the accuracy of the cumulative consumption by accumulating them can be ensured to some extent, so a prediction method that does not use day-of-week sensitivity data is also possible It is. Even when day-of-week sensitivity data is used, the day-of-week sensitivity data is not limited to the above calculation method.

  Furthermore, the calculation method of the daily temperature correction coefficient is not limited to the above-described calculation method. The daily temperature correction coefficient does not necessarily have to be calculated for each customer. For example, the same daily temperature correction coefficient may be used for each customer, or may be calculated for each customer region.

  Further, the customer information database DB1 or the cylinder information database DB2 stores gas consuming equipment information such as the type and number of gas consuming equipment 22 (see FIG. 2) connected to the LPG cylinder 21 of each customer. When the fuel consumption amount predicting means 11 is changed in the gas consuming equipment information when the number of units is changed, the fuel consumption amount predicting means 11 is changed when the fuel consumption amount information is changed. It is also preferable that the predicted value of the daily gas consumption is corrected based on the change content.

<4> In the above embodiment, it is assumed that the fuel usage amount prediction means 11 is provided as a part of the system 1 of the present invention, and the daily gas consumption for each customer is predicted when the mathematical plan problem of the delivery plan is derived. Using the fuel usage amount predicting means provided separately from the system 1 of the present invention, the predicted value of the daily gas consumption of each customer in a predetermined period including at least the gas consumption calculation period is calculated. It may be stored in the consumption database DB7, and the mathematical plan solution means 12 may read the daily gas consumption prediction value from the fuel consumption database DB7 and use it as necessary. In this case, step # 4 shown in FIG. 3 is not performed. The calculation of the predicted value of the daily gas consumption may be performed, for example, every time the actual value of the daily gas consumption is updated.

<5> In the above embodiment, the shortest travel time for each date and time stored in the travel time database DB6 is calculated in advance using the passage time information of each road calculated in advance and stored in the road information database DB5. However, whenever the values of the constants A1 _{t, i, j, and δ} of Equations 4 and 15 are required without providing the travel time database DB6, they correspond to A1 _{t, i, j, and δ} . The shortest travel time may be calculated each time. Further, the passage time information may be calculated each time based on the standard passage time and the passage time correction value stored in the road information database DB5.

<6> In the above embodiment, the case where there is one delivery base has been described. However, a plurality of delivery bases may be provided. In addition, a delivery base and a customer house managed by the delivery base may be linked in advance. In addition, when the delivery base and the customer's house correspond one-to-one and the delivery base and the worker correspond one-to-one, there are as many mathematical programming problems as there are delivery bases in the number of delivery bases. become.

<7> Further, in the above-described embodiment, it is assumed that the worker does not return to the delivery base in the middle of the daily fuel delivery work, but may be allowed to return to the delivery base in the middle. In this case, it can be dealt with by correcting the above-mentioned delivery amount constraint condition and the constraint condition between the variables γ _{t, i, j, k} .

<8> In the above embodiment, the constraint condition for the variable α _{t, i} is not particularly set. For example, within the planning target period that is normally set, the number of fuel delivery operations to each customer's house is 1 or more. Since it is 3-5 or less at most, it is also preferable to add the constraint condition shown in the following equation (26). The right side N (T) of Expression 26 represents the maximum value of the number of times of fuel delivery work determined by the number of days T in the planning target period.

<9> In the above embodiment, it is assumed that the LPG cylinders individually installed in the customer's house are exchanged and collected as the fuel delivery work from which the delivery plan is derived by the system 1 of the present invention. On the other hand, instead of the replacement and collection, an operation of replenishing LPG in the LPG cylinder may be performed. Further, the fuel to be delivered by the fuel delivery operation is not limited to LPG, and any fuel such as light oil or kerosene that can supply the fuel by the delivery operation by the operator is delivered to the fuel. The plan is to be derived by the system 1 of the present invention. Also, the type of fuel is not limited to one, and a plurality of fuels may be mixed. In this case, the fuel usage amount predicting means 11 predicts the daily fuel consumption for each type of fuel.

<10> In the above embodiment, as the work constraint conditions, the above-mentioned work condition items (standard working days per month, standard working hours for one day, continuous working days, continuous working days spanning from the end of the month to the next month, We assumed a restriction condition for holidays within the planning target period that we wanted to acquire), but in addition to or in place of the relevant work conditions item, we could use multiple months or years of work days good.

  The delivery load leveling system according to the present invention is configured to carry out operations such as replacement and recovery of fuel containers or refueling of fuel containers, from a delivery base by mounting a replacement fuel container or fuel for replenishment on a work vehicle. It is a delivery plan for fuel delivery work that goes around the customer's home and can be used to derive a delivery plan that equalizes the work load of multiple workers over a predetermined planning period by computer processing. is there.

1: Distribution load leveling system 2: Arithmetic processing device 3: Storage device 4: Communication device 11: Fuel consumption predicting means 12: Mathematical plan solving means 21: LPG cylinder 22: Gas consuming equipment 23: Adjuster 24: Gas meter 25 : Data totaling device 26: Communication device 27: Communication line 28: Storage device 29: Information processing terminal 30: Wireless data communication line DB1: Customer information database DB2: Cylinder information database DB3: Worker information database DB4: Air temperature information database DB5: Road information database DB6: Travel time database DB7: Fuel consumption database

Claims (10)

Replacing and collecting fuel containers individually installed at a plurality of customer homes or performing customer home work such as refueling the fuel containers, mounting the fuel containers for replacement or fuel for replenishment on work vehicles, A delivery plan for a fuel delivery work performed by patroling the customer premises from a delivery base, wherein a delivery plan in which the work load of a plurality of workers is leveled over a predetermined target period is derived by computer processing A load leveling system,
A fuel container information database, an operator information database, an air temperature information database, and a road information database are formed in a storage area constituted by one or a plurality of storage devices, and mathematical operations are performed in the arithmetic processing unit constituting the computer. Planning solution means are configured,
The fuel container information database includes location information of the customer's home and installation information of the fuel container including attribute information of the installed fuel container, a delivery work date when the fuel delivery work is performed on the fuel container, and The fuel supply information of the fuel container including the remaining amount of the fuel container before the fuel delivery work is stored so as to be searchable for each customer.
The worker information database stores work-related information including standard work days of at least a month and standard work hours of one day for the worker in the planning target period so as to be searchable for each worker.
The temperature information database stores temperature information within at least the planning target period so as to be searchable by region of the installation location of the fuel container,
At least road time information required for the road information database to pass through each road in the work area including the delivery base and all the customer houses, and road information necessary to generate the passage time information. Either one is stored so that it can be searched by date and by road,
The mathematical program solving means is
The installation information and the fuel supply information read from the fuel container information database, the work related information read from the worker information database, the temperature information read from the temperature information database, read from the road information database, or Accepting as input information the transit time information generated based on the road information read from the road information database, and the deliverable amount that the work vehicle can deliver in one fuel delivery operation;
Based on the input information, the remaining amount restriction condition for the remaining amount of the fuel container in all the customer's homes not to disappear, the working condition items relating to working days and working hours for each worker during the planning period Work constraint conditions for satisfying the standard work conditions included in the work-related information, and the fuel delivery amount for all of the fuel containers allocated to one fuel delivery operation is less than the deliverable amount Set the delivery amount constraint condition as
Based on each constraint condition, with the remaining amount constraint condition and the delivery amount constraint condition as absolute conditions, the work condition item for each worker in the work constraint condition deviates from the corresponding standard work condition. The fuel container to be the target of the fuel delivery work for each worker on each work day during the planning period and the work order so that the objective function for summing up the first penalty value indicating the degree of A delivery load leveling system configured to solve the delivery plan including a mathematical planning problem. The mathematical program solution means includes a fuel consumption amount prediction means,
The fuel usage amount prediction means reads the installation information and the fuel supply information from the fuel container information database, and the temperature information from the temperature information database, respectively, and based on the installation information, the fuel supply information, and the temperature information. The delivery load leveling according to claim 1, wherein a remaining amount of the fuel container is predicted for each of the fuel containers on a date provisionally set during the solution process of the mathematical programming problem. System.   The mathematical plan solving means uses the fuel consumption predicting means for each customer, on a day when there is no remaining amount of the fuel container generated during the planning target period and within a predetermined grace period after the planning target period has elapsed. One or a plurality of limit dates, and for each customer, one or a plurality of delivery work set dates for performing the fuel delivery work set during the solving process of the mathematical programming problem are set as the delivery work set dates. 3. The delivery load leveling system according to claim 2, wherein the delivery load leveling system is set within a recommended delivery period within the grace period from the limit date before the limit date corresponding in the same order.   The mathematical plan solution means, when the delivery work setting date is not set for each customer within a second recommended delivery period within a second grace period shorter than the grace days from the limit date before the limit date. The second penalty value indicating the degree of deviation from the second recommended delivery period of the delivery work setting date is calculated, and instead of an objective function for counting the first penalty value, the first penalty value and the first penalty value are calculated. 4. The delivery load leveling system according to claim 3, wherein the delivery plan is derived so that an objective function for totaling two penalty values is minimized.   The delivery load leveling according to claim 4, wherein an upper limit value is set individually or total for at least one of the first penalty value and the second penalty value. System.   6. The delivery load level according to claim 4, wherein the second penalty value increases further as a deviation from the second recommended delivery period becomes larger than a value proportional to the degree of the deviation. System.   7. The first penalty value increases as the deviation from the standard working condition increases further than a value proportional to the degree of the deviation. 8. Delivery load leveling system. A travel time database formed in the storage area;
The road information database includes the shortest travel time along a movable route between one customer house and the other customer house and the delivery base existing within a predetermined travel distance from the one customer house. Based on the transit time information or the road information stored in, for each date and time within the planning target period, it is calculated in advance for all the customers, and is stored in the travel time database,
The delivery load leveling system according to any one of claims 1 to 7, wherein the mathematical plan solving means uses the shortest travel time instead of the transit time information as the input information.   The mathematical plan solving means circulates the time required for one fuel delivery operation in the order set from the delivery base through the installation locations of the fuel containers set for the one fuel delivery operation. The total travel time until returning to the delivery base and the total work time of the customer home work at each of the installation locations of the fuel container are calculated at least in total. The delivery load leveling system according to item 1.   The said working condition item contains the working days of a month, the working hours of a day, the continuous working days in the same month, and the continuous working days over a month, In any one of Claims 1-9 characterized by the above-mentioned. The described delivery load leveling system.