JP2006331340A - Traffic line processing device, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To output traffic line data, which hold a main traffic line acquired from actual measurement data, as a simulation result. <P>SOLUTION: First and second main traffic lines are extracted from a set of actual measurement traffic lines and a set of simulation traffic lines. From the simulation traffic lines larger in number than a value found by subtracting the number corresponding to a main traffic line including ratio from the number of simulation traffic lines, factors not included in the respective first main traffic lines are removed from those of the respective second main traffic lines. The simulation traffic lines each including the first main traffic lines are classified into the same group. From the simulation traffic lines in the respective groups, the first main traffic lines except the corresponding first main traffic lines are eliminated. From a set of simulation traffic lines after elimination, traffic lines as many as those lacked in the group are selected, and the corresponding first main traffic lines are inserted. From the set of the simulation traffic lines after insertion, third main traffic lines are extracted. If a traffic line except the set of the first traffic lines is included in a set of the third traffic lines, the traffic line is eliminated from the simulation traffic line including the main traffic line. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a flow line processing apparatus, method, and program.

  The model of human behavior in a specific space is emphasized in the fields of architecture, civil engineering, and marketing, and various studies have been conducted. Commercial facilities such as retail stores and theme parks are also one of the target spaces. . Conventional modeling methods are roughly divided into the following two.

  One is based on cellular automata, and is modeled after retail stores and theme parks.

On the other hand, there is a probabilistic utility maximization model that uses individual and field attributes as explanatory variables and selects the most effective one from a plurality of options. Applications of the probabilistic utility maximization model include consumer purchase selection and traveler means selection in traffic behavior.
Japanese Patent Application No. 2002-208809 Japanese Patent Application No. 2001-5016 “Ubiquitous information based in-store excursion model” (author: Ichiro Toyoshima, Takashi Komine, Atsushi Yoshida, Kanako Hattori, Naoki Imazaki, publication: IEICE technical report Vol.104 No.727 “Artificial intelligence and knowledge Processing ”, issue date: March 8, 2005, ISSN: 0913-5685)

  The conditions to be satisfied by the daily migratory behavior model in a commercial facility reflect the attributes of individual customers and places in the behavior prediction, and the overall behavior in the store, for example, arrival time, stay time, target goods selection, route selection However, the cellular automaton type model cannot cope with the former, and the probability utility maximization model cannot cope with the latter. Further, in the conventional model, since the estimation using the measurement data of the movement history (flow line) is not performed, the reliability of the simulation result is low.

  Therefore, in the Non-Patent Document 1, the present inventors describe the relationship between individual attributes and arrival time intervals, movement probabilities between sales floors, and stay time / target spot selection / purchase determination / stay end determination. We proposed a method for predicting customer behavior in a store on the day by estimating data and measured data of movement history and inputting virtual personal attributes.

  However, the customer's movement history obtained by this method may not hold the main flow line included in the measured data of the movement history.In this case, effective layout and personnel allocation based on the simulation results Have difficulty.

  The present invention provides a flow line processing apparatus, method, and program capable of providing flow line data holding main flow lines obtained from measured data as simulation results.

  The flow line processing apparatus according to one aspect of the present invention is a measured flow line in which a simulation flow line set, which is a set of flow lines of a moving body obtained by simulation, is a set of flow lines of a moving body, which is obtained by actual measurement. A flow line processing apparatus that adjusts the simulation flow line set so as to hold a main flow line extracted from the set, the first input means for inputting the measured flow line set, and the simulation flow line set being input Second input means, first main flow line extracting means for extracting a first main flow line from the measured flow line set according to a pre-given main flow line content rate, and the simulation flow line according to the main flow line content rate In the case where the second main flow line extracting means for extracting the second main flow line from the set and the second main flow line set and the first main flow line set do not coincide with each other, A determining means for determining an element that is not included in each of the first main flow lines, and the number of simulation flow lines corresponding to the main flow line content rate from the number of simulation flow lines included in the simulation flow line set. For each of the first main flow lines, a removal means for selecting a number of simulation flow lines larger than the value obtained by subtracting from the simulation flow line set and removing the determined element from the selected simulation flow lines, Classifying means for detecting a simulation flow line including the first main flow line from a set of simulation flow lines that have passed through the removing means, and classifying the detected simulation flow lines into the same group, and a group selection criterion given in advance From the simulation flow line in the group according to the order of the group based on The main flow line deleting means for deleting the first main flow line other than the first main flow line corresponding to the group, and for each of the groups, the number of simulation flow lines included in the group is the main flow line content rate. In order to satisfy, from the simulation flow line that does not include any of the first main flow lines in the simulation flow line set that has undergone the main flow line removal means, the number of simulation flow lines that are insufficient in the group is obtained. A main flow line inserting unit that selects and inserts a first main flow line corresponding to the group into the selected simulation flow line, and a third main movement according to the main flow line content rate from the simulation flow line set that has passed through the main flow line insertion unit. If a third main flow line extracting means for extracting a line and a main flow line other than the first main flow line set are included in the third main flow line set, Comprising a further main drive line deleting means for deleting the main drive line from the simulation flow line comprising main drive line, and an output means for outputting a simulation flow line set having passed through the main driving line removal means.

  In the method according to one aspect of the present invention, a simulation flow line set that is a set of flow lines of a moving object obtained by simulation is estimated from an actual flow line set that is a set of flow lines of a moving object obtained by measurement. A method of adjusting the simulation flow line set so as to hold the main flow line, wherein the first input step inputs the actual measurement flow line set and the second input step inputs the simulation flow line set. A first main flow line extraction step for extracting a first main flow line from the measured flow line set according to a pre-given main flow line content rate; and a second main movement from the simulation flow line set according to the main flow line content rate If the second main flow line extracting step for extracting a line does not match the second main flow line set and the first main flow line set, Of the elements, a determination step for determining an element not included in each of the first main flow lines, and the number of simulation flow lines corresponding to the main flow line content rate from the number of simulation flow lines included in the simulation flow line set. For each of the first main flow lines, a removal step of selecting more simulation flow lines than the value obtained by subtracting from the simulation flow line set and removing the determined element from the selected simulation flow lines; A classification step of detecting a simulation flow line including the first main flow line from the simulation flow line set after the removal step, and classifying the detected simulation flow lines into the same group, and a group selection criterion given in advance The group order based on the simulation within the group A main flow line removing step of deleting a first main flow line other than the first main flow line corresponding to the group from the movement line, and for each of the groups, the number of simulation flow lines included in the group is the main movement In order to satisfy the line content rate, from the simulation flow line that does not include any of the first main flow lines in the simulation flow line set after the main flow line removal step, A main flow line insertion step of selecting a simulation flow line and inserting a first main flow line corresponding to the group into the selected simulation flow line, and from the simulation flow line set after the main flow line insertion step according to the main flow line content rate A further main flow line extracting step for extracting a third main flow line, and a set of the first main flow lines in the third main flow line set When an external main flow line is included, a main flow line deletion step for deleting the main flow line from the simulation flow line including the main flow line, and an output step for outputting a simulation flow line set after the main flow line deletion step, Prepare.

  The program according to one aspect of the present invention estimates a simulation flow line set, which is a set of mobile flow lines obtained by simulation, from a measured flow line set, which is a set of mobile flow lines obtained by measurement. A program for adjusting the simulation flow line set so as to hold the main flow line thus generated, a first input step for inputting the actual measurement flow line set, and a second input for inputting the simulation flow line set An input step; a first main flow line extracting step for extracting a first main flow line from the measured flow line set according to a pre-given main flow line content rate; and a second from the simulation flow line set according to the main flow line content rate. If the second main flow line extracting step for extracting the main main flow line and the second main flow line set do not match the first main flow line set, Corresponding to the main flow line content rate from the determination step of determining elements not included in each of the first main flow lines among the elements in the two main flow lines and the number of simulation flow lines included in the simulation flow line set Selecting a simulation flow line having a number greater than a value obtained by subtracting the number of simulation flow lines to be performed from the simulation flow line set, and removing the determined element from the selected simulation flow line; A classification step of detecting a simulation flow line including the first main flow line from the simulation flow line set after the removal step and classifying the detected simulation flow lines into the same group for each of the main flow lines; According to the group order based on the given group selection criteria, the group A main flow line removing step of deleting a first main flow line other than the first main flow line corresponding to the group from the simulation flow line in the group, and the number of simulation flow lines included in the group for each of the groups In order to satisfy the main flow line content rate, the group is insufficient from the simulation flow line that does not include any of the first main flow lines in the simulation flow line set after the main flow line removal step. A main flow line inserting step of inserting a first main flow line corresponding to the group into the selected simulation flow line, and the main flow line from the simulation flow line set after the main flow line insertion step. A further main flow line extracting step for extracting a third main flow line according to the content rate; If a main flow line other than the set of one main flow line is included, a main flow line deletion step for deleting the main flow line from the simulation flow line including the main flow line, and a simulation flow line set after the main flow line deletion step are output. And causing the computer to execute an output step.

  According to the present invention, the flow line data that holds the main flow line obtained from the measured data can be given as a simulation result.

  FIG. 1 is a block diagram schematically showing the configuration of the behavior prediction apparatus according to the present invention.

  This behavior prediction apparatus includes an attribute information database (DB) 11, an actual measurement data acquisition unit 12, a moving object generation interval estimation unit 13, a target area estimation unit 14, an area stay time estimation unit 15, an inter-area transit time estimation unit 16, a movement A path estimation unit 17, a flow line generation condition setting unit 18, a flow line simulation unit (flow line processing device) 19, a main flow line extraction unit 20, a flow line adjustment unit 21, and a flow line data storage unit 22 are provided.

  The attribute information database 11 holds a plurality of attributes for each moving object. The attributes include, for example, gender, number of people, number of customers visiting the store, and a target area (target sales floor) of the moving object. Information on these attributes is acquired by a questionnaire or the like for customers who visit the store.

  The actual measurement data acquisition unit 12 acquires and holds movement history data (flow line data) of customers who have visited the store. The movement history is acquired using, for example, RFID. That is, the movement history is acquired by having the customer hold the RFID and placing the receiver in the store. The flow line data includes the order in which the moving body visited each area. In addition, the flow line data also includes information such as store entry time, stay time in each area, and store exit time.

  The moving body generation interval estimation unit 13 receives as input the attribute information in the attribute information database 11 and a set of flow line data (flow line set) acquired by the actual measurement data acquisition unit 12. The moving body generation interval estimation unit 13 uses these input data to estimate a parameter of a relational expression between the personal attribute and the moving body generation interval as a parameter such as an acceleration / failure model, and outputs the estimated parameter.

  FIG. 2 is a block diagram showing the configuration of the moving object generation interval estimation unit 13.

  The moving body generation interval estimation unit 13 includes an attribute-specific generation time interval extraction unit 13a and a moving body generation interval model parameter estimation unit 13b.

  The attribute-specific occurrence time interval extraction unit 13a extracts mobile objects having the same combination of attributes from the attribute information database 11, acquires the arrival times of the extracted mobile objects from the flow line set, respectively (attribute 1, attribute 2... Information such as attribute m, arrival time, arrival time,... Arrival time) is generated. Attribute 1, attribute 2,... Attribute m in parentheses are a combination of certain attributes. There are as many arrival times in parentheses as the number of extracted mobiles. This information is generated for the number of attribute combinations. You may specify the attribute used in the estimation from the attribute setting part 13c (not shown in FIG. 1).

  The moving body generation interval model parameter estimation unit 13b receives the information generated for the number of attribute combinations as input, and performs parameter estimation of the moving body generation interval model. This moving object generation interval model takes personal attributes as input and outputs a probability distribution of moving object generation intervals (customer arrival intervals). The moving body generation interval model parameter estimation unit 13b outputs the estimation result as a moving body generation interval model coefficient estimated value vector.

  Returning to FIG. 1, the target area estimation unit 14 receives the attribute information in the attribute information database 11 and estimates a parameter of a relational expression between a personal attribute and a target area selection probability as a parameter of a multinomial logit model, Output the estimated parameters.

  FIG. 3 is a block diagram illustrating a configuration of the target area estimation unit 14.

  The target area estimation unit 14 includes a target area model parameter estimation unit 14a.

  The target area model parameter estimation unit 14a reads the attribute information (including the target area) of each moving object from the attribute information database 11, and estimates, for example, the parameters of the multinomial logit model from the read attribute information of each moving object. In this multinomial logit model, when a personal attribute is received as an input, the selection probability of each alternative (area) is calculated, and the alternative having the largest value is output as a decision (target area). The target area model parameter estimation unit 14a outputs the estimation result as a target area model coefficient estimated value vector. You may specify the attribute used in the estimation from the attribute setting part 14b (not shown in FIG. 1).

  Returning to FIG. 1, the area stay time estimation unit 15 receives the attribute information in the attribute information database 11 and the flow line set acquired by the actual measurement data acquisition unit 12 as input parameters such as an acceleration / failure model and the like. The parameter of the relational expression between the attribute and the stay time distribution in each area is estimated, and the estimated parameter is output.

  FIG. 4 is a block diagram showing a configuration of the area stay time estimation unit 15.

  Area stay time estimation unit 15 includes area-specific / attribute-specific stay time extraction unit 15a, stay time model parameter estimation unit 15b (area 1 stay time model parameter estimation unit 15b-1, area 2 stay time model parameter estimation unit 15b-2. ,... Has an area e staying time model parameter estimation unit 15b-e).

  The area-specific / attribute-specific stay time extraction unit 15a identifies a mobile object that includes the area in the flow line data for each area, divides the identified mobile object into groups having the same attribute combination, and for each group, the group The staying time of each moving body included in is acquired from the flow line set. Then, information such as (attribute 1, attribute 2,... Attribute m, stay time, stay time,... Stay time) is generated for each area by the number of groups. Attribute 1, attribute 2,... Attribute m in parentheses are combinations of attributes. There are as many staying times in parentheses as there are mobiles included in the group. You may specify the attribute used in the estimation from the attribute setting part 15c (not shown in FIG. 1).

  The area 1 stay time model parameter estimation unit 15b-1 uses the information of area 1 to estimate the parameters of the model having the personal attributes as input and the stay time distribution in area 1 as output, and the estimation result is used as the area 1 stay. Output as a time model coefficient estimate vector.

  Similarly, the area 2 stay time model parameter estimation unit 15b-2 to the area e stay time model parameter estimation unit 15b-e use the information of the area 2 to the area e and input the personal attribute to input the area 2 to the area e. The parameter of the model that outputs the stay time distribution is estimated, and the estimation result is output as an area 2 stay time model coefficient estimated value vector to an area e stay time model coefficient estimated value vector.

  Returning to FIG. 1, the inter-area transit time estimation unit 16 receives the flow line set obtained by the actual measurement data acquisition unit 12 as an input, and is a model parameter representing a travel time index (for example, travel time probability distribution) between the areas. Is estimated for each area, and the model parameters of each estimated area are output.

  FIG. 5 is a block diagram showing a configuration of the inter-area transit time estimation unit 16.

  The inter-area transit time estimation unit 16 includes a route stay time extraction unit 16a, an inter-route stay time model parameter estimation unit 16b (route (area 1, area 2) stay time model parameter estimation unit 16b-1, and a route (area 1). , Area 3) inter-stay time model parameter estimation unit 16b-2,... Route (area 2, area 1) inter-stay time model parameter estimation unit 16b-x,... Route (area e-1, area e) Inter-stay time model parameter estimation unit 16b-y).

  The route stay time extraction unit 16a extracts all the travel time of the route for each route (between areas) from the flow line set acquired by the actual measurement data acquisition unit 12, and the route (first area, second area). : (Travel time, travel time,... Travel time) is generated for the number of types of existing routes. There are as many travel times in parentheses as there are routes (between the first area and the second area).

  The stay time model parameter estimation unit 16b-1 between the routes (area 1, area 2) calculates the area 1 and area from the information of the route (area 1, area 2): (travel time, travel time,... Travel time). The parameter of the model representing the probability distribution of the travel time between the two is estimated, and the estimation result is output as a route (area 1, area 2) stay time model parameter estimated value vector.

  Similarly, the stay time model parameter estimation unit 16b-2 between the routes (area 1, area 3) to the stay time model parameter estimation unit 16b-y between the routes (area e-1, area e) perform the same estimation. Route (area 1, area 3) stay time model parameter estimated value vector to route (area e-1, area e) stay time model parameter estimated value vector are output.

  Returning to FIG. 1, the movement path estimation unit 17 takes the flow line set obtained by the actual measurement data acquisition unit 12 as an input, estimates the parameters of the movement probability model for each area as a parameter such as a hidden Markov model, Output parameters for each estimated area. This model outputs the generation probability of each route between areas for a route equal to or shorter than the length of the route length parameter from the source area, the destination area, the route length parameter, and the like. In actual route selection, for example, a route having a high generation probability (for example, the top three) is adopted as a candidate, and the route is selected at random after the adopted candidate is weighted by the generation probability.

  FIG. 6 is a block diagram showing a configuration of the movement route estimation unit 17.

  The movement route estimation unit 17 includes a route use frequency calculation unit 17a for each movement source / target area, a route selection model parameter estimation unit 17b (route (area 1, area 2) route selection model parameter estimation unit 17b-1, a route ( Area 1, area 3) route selection model parameter estimation unit 17b-2 ... route (area 2, area 1) route selection model parameter estimation unit 17b-x ... route (area e-1, area e) And an inter-path selection model parameter estimation unit 17b-y).

  The route usage frequency calculation unit 17a for each movement source / target area extracts all the routes for each area from the flow line set acquired by the actual measurement data acquisition unit 12, and the route (the first area, the first area). 2 area): information such as (route, route,... Route) is generated between the areas. Paths in parentheses may overlap, that is, when a plurality of the same paths are included in the flow line set, a plurality of the same paths exist in the parentheses.

  The route selection model parameter estimation unit 17b-1 between routes (area 1, area 2) calculates the area from the information of route (area 1 (a1), area 2 (a2)): (route, route,... Route). The parameters of the movement probability model between 1 and area 2 are estimated, and the estimation result is output as a route (area 1, area 2) route selection model coefficient estimated value vector.

  Similarly, the route selection model parameter estimation unit 17b-2 to the route (area e-1, area e) between routes (area 1, area 3) also performs parameter estimation. The route (area 1, area 3) route selection model coefficient estimated value vector to route (area e-1, area e) route selection model coefficient estimated value vector are output.

  Returning to FIG. 1, the flow line generation condition setting unit 18 sets conditions (flow line generation conditions) for performing simulation in the flow line simulation unit 19. Specifically, each data (attribute 1, attribute 2,...) Of the moving bodies 1 to n to be simulated is set.

  The flow line simulation unit 19 performs a simulation using the parameter group output by each of the estimation units 13 to 17 and the flow line generation condition data set by the flow line generation condition setting unit 18 as input, and performs the movement of each moving body. Output line data (flow line set).

  FIG. 7 is a block diagram showing a configuration of the flow line simulation unit 19.

  The flow line simulation unit 19 includes an arrival time generation unit 19a, a stay time generation unit 19b, a target area generation unit 19c, and a movement route generation unit 19d.

  The arrival time generation unit 19a determines the arrival time of each mobile unit from the mobile unit generation interval model coefficient estimated value vector output by the mobile unit generation interval estimation unit 13 and the flow line generation condition data.

  The stay time generation unit 19b determines the stay time in each area of each moving body from each area stay time model coefficient estimated value vector output by the area stay time estimation unit 15 and the flow line generation condition data.

  The target area generation unit 19c determines the target area of each moving object from the target area model coefficient estimated value vector output by the target area estimation unit 14 and the flow line generation condition data.

  The movement path generation unit 19d determines the movement time between the areas based on the inter-area passage time model coefficient estimated value vector output by the inter-area passage time estimation unit 16. Note that this process may be performed whenever necessary during the process of the movement path generation unit 19d. Further, the movement route generation unit 19d determines a movement route from the movement source area and the target area based on each route selection model coefficient estimated value vector output by the movement route estimation unit 17.

  FIG. 8 is a flowchart for explaining in detail an example of processing by the movement route generation unit 19d.

  First, when a movement route to the target area is generated (S1), the exit determination area is set as the start area (S2). The exit determination area is advanced by one according to the generated travel route (S3). The elapsed time from the start point area to the current point area is calculated from the stay time in each area determined by the stay time generator 19b and the travel time between the areas (S4). When the stay end determination is made (S5) and the stay is ended (for example, when the elapsed time from entering the store exceeds the planned stay time) (YES in S5), the route selection model (each route selection model coefficient estimated value vector) ), The exit route is determined (S6). On the other hand, if the stay is not finished as a result of the stay end determination (NO in S6), it is determined whether or not the exit determination area matches the target area (S7). In the stay end determination, a model for stay end determination may be generated in advance and the stay end determination may be performed based on this model. If the exit determination area does not match the target area (NO in S7), the process returns to step S3. On the other hand, if the exit determination area does not coincide with the target area (YES in S7), the next target area is generated using the target area generating unit 19c (S8), and the process returns to step S1.

  By performing the above processing for each moving body, as shown in FIG. 7, flow line data (flow line set) of each moving body is generated. The generated flow line set is transferred to the flow line adjusting unit 21 in FIG.

  In FIG. 1, the main flow line extraction unit 20 extracts a main flow line based on the flow line data of each moving body acquired by the actual measurement data acquisition unit 12. The main flow line is an ordered flow line pattern that exceeds a predetermined content rate in all flow line data.

  The flow line adjustment unit 21 sets the number of movement lines in the flow line set so that the flow line data (flow line set) of each moving body received from the flow line simulation unit 19 holds the main flow line extracted by the main flow line extraction unit 20. The flow line data is adjusted, and the adjusted flow line set is stored in the flow line data storage unit 22 as a simulation result. That is, the flow line adjustment unit 21 functions as a post-process of the flow line simulation unit 19. More specifically, the process by the flow line adjustment unit 21 is performed by the flow line data conversion unit 21a as shown in FIG.

Hereinafter, processing by the flow line data conversion unit 21a will be described in detail with reference to the drawings.
First, symbols used in the description are defined.

(1) S: Spot (area) set
S = {s ₁₁ , ... s _n }. Each s _i is "department", corresponds to the spot (observation point), such as "doorway".

(2) T: Customer flow line set The flow line set of all customers is expressed as T = {t ₁ ,... T _N }. The flow line of each customer is expressed as t _i = s _i1 s _i2 ... S _ik (where each s _i is a spot), and t _i corresponds to the travel route per customer. For example, when the customer moves from entrance → TV section → refrigerator section → exit, each point corresponds to s.

(3) mt: main flow line As described above, an ordered flow line pattern that exists in a flow line set in excess of a predetermined content rate. The notation is mt _i = {s _i1 , s _i2 ,..., S _ij }. A specific example of the main flow line is shown in FIG. Circled numbers 1 to 11 represent spots. Here, the content P is set to 0.25. The patterns need only match the order of the spots, and the spots need not be continuous. In addition, a plurality of main flow lines may be included in one customer flow line.

(4) MT: main flow line set The main flow line set is a set of main flow lines described in (3). MT = {mt ₁ , mt ₂ , ... mt _j }. Hereinafter, a set of main flow lines obtained from the measured data is denoted as MT, and a set of main flow lines obtained from the flow line set of the simulation results is denoted as MT ′. An example of the main flow line set MT is shown in FIG.

(5) SP: Set of main flow line components The set of main flow line components is an unordered set of spots that make up the main flow line. Expressed as SP (mt). An example of the main flow line component set SP is shown in FIG.

と表記する。NUの一例を図１３に示す。 (6) NU: A set of spots that are not included in the main flow line obtained from the measured data, but are included in the main flow line obtained from the simulation result set

Is written. An example of NU is shown in FIG.

(7) mtεt: the main flow line mt is included in the flow line t That is, mtεt indicates a state in which all elements of mt exist in the flow line t in a state satisfying the order. An example where mtεt is satisfied and an example where mtεt is not satisfied are shown in FIG.

(8) L (mt): A set of flow lines including the main flow line mt
L (mt) = {t | mt∈t}. On the other hand, let L _no be a set of flow lines that do not include any main flow line.

  FIG. 15 is a flowchart for explaining processing by the flow line data conversion unit 21a.

  First, the flow line data conversion unit 21a uses, as input data, a flow line set obtained by the flow line simulation unit 19, a main flow line content rate, and a main flow line set obtained from actually measured data. An example of these data used in the following description is shown in FIG.

The flow line data conversion unit 21a obtains the main flow line set MT ′ from the customer flow line set T = {t ₁ ,... T ₁₀ } by the flow line simulation unit 19. It is determined whether MT ′ and MT match (S11). If they match (YES in S11), the customer flow line set T is output as a processing result and stored in the flow line data storage unit 22. .

  On the other hand, when MT ′ and MT do not match (NO in S11), NU is determined (S12). This is shown in FIG. The processing result is NU = {spot 9}.

Next, N− (N × p) flow lines are randomly selected from the customer flow line set T, and NU elements are removed from the selected flow lines (S13). N is the number of elements of T. FIG. 18 shows how this step is performed. Since N− (N × p) = 10− (10 × 0.25) ≒ 8 (round up 0.75), select 8 flow lines at random (here, select flow lines other than t ₃ and t ₈₎ ), The spot 9 is removed from the selected eight flow lines. This ensures that the spot 9 does not appear in the main flow line obtained from the customer flow line set.

Next, L (mt _i ) is obtained for all i (Classfy (Li)) (S14). That is, the flow lines in the customer flow line set are classified into L (mt _i ). This is shown in FIG. When a certain flow line includes two or more main flow lines, the flow line belongs to a plurality of categories. Here, t ₃ is an example.

Next, in order from L (mti) with a small number of elements, each element in L (mti) is operated on the corresponding element (flow line) so as to include only the corresponding main flow line (Fix (Li)) ( S15). That is, one flow line does not include a plurality of main flow lines. The state of this operation is shown in FIG. Since L (mt ₂ ) has fewer elements than L (mt ₁ ), the operation is performed from L (mt ₂ ). By performing the operation from L (mt ₂ ) having a small number of elements in this way, the calculation amount in this step can be reduced. The order in which operations are performed may be determined randomly. The overlap of the main flow line included in the flow line t ₃ disappears by the operation on L (mt ₂ ). As a result, an operation for L (mt ₁ ) becomes unnecessary (see FIG. 19). A specific example of processing for t ₃ is shown in FIG. First, all the elements of MT existing in the flow line t ₃ are once deleted, and then the corresponding main flow line mt ₂ (spot 4−spot 8) is inserted at a random position in t ₃ (Set (mt ₂ )). Here, the spot 4 and the spot 8 are inserted in a continuous state. However, as long as the order of the spots is maintained, these spots may be inserted in a state of being separated across the spots 9. As described above, the flow line t ₃ is guaranteed to belong only to L (mt ₂ ).

Then, L (mt _i) the number of elements is the content of (p = 0.25) by operating the customer flow line set to exceed the increase the number of elements _{L (mt i) (Raise (} Li)) (S16 ). In this example, since p = 0.25, the number of elements of L (mt _i ) is increased to at least three. The operation may be performed from L (mt _i ) having a small number of elements, and in this case, the calculation amount in this step can be reduced. Subject to flow-line operations randomly selected from flow line in L _no, applying the Set (mti) for the selected flow line. By selecting a flow line from L _{no, the} possibility that an unnecessary main flow line is extracted in the next step S16 can be reduced. It is preferable that the flow lines to be selected do not overlap in each mt _i . The state of the operation in this step is shown in FIG. For example, in the case of L (mt ₁ ), it is assumed that two flow lines are selected at random, and as a result, flow lines t2 and t8 are selected. All elements of MT are once deleted from the selected flow line t2 (here, spot 4 and spot 2 are deleted), and then the corresponding main flow line (spot 1-spot 2-spot 3) is randomly positioned. (Set (mt ₁ )). The main flow spot to be inserted does not need to be continuous as long as the order is maintained. The same applies to the flow line t8.

  Next, the main flow line extraction algorithm is applied to the customer flow line set after step S16 to extract the main flow line, and the main flow line set MT 'is updated with the extracted main flow line (S17). FIG. 23 shows how this processing is performed. It is guaranteed that the two required main flow lines (= elements of MT) are extracted by steps S12 to S16 described above. Here, other main flow lines (spot 2 -spot 4 -spot 3) are also extracted, which are unnecessary main flow lines (Restmt).

  Next, in the customer flow line set, the main flow line is deleted from the flow lines including unnecessary main flow lines (Remove Rest (mti)) (S18). FIG. 24 shows how this processing is performed. Since unnecessary main flow lines are included in the flow lines t6, t7, and t9, unnecessary main flow lines are deleted from these flow lines. The customer flow line set after deletion is shown in FIG. This flow line set becomes the output of the flow line adjustment unit 21 and is stored in the flow line data storage unit 22. This customer flow line set coincides with the main flow line obtained from the actual measurement data.

  As described above, according to the present embodiment, a set of flow line data that retains the reproducibility of the main flow line obtained from measured data, that is, a set of flow line data that reflects the customer's route characteristics is generated. Can do. This makes it possible to propose an effective layout and personnel arrangement.

It is a block diagram which shows roughly the structure of the action prediction apparatus concerning this invention. It is a block diagram which shows the structure of a moving body generation | occurrence | production space | interval estimation part. It is a block diagram which shows the structure of a target area estimation part. It is a block diagram which shows the structure of an area stay time estimation part. It is a block diagram which shows the structure of the passage time estimation part between areas. It is a block diagram which shows the structure of a movement path | route estimation part. It is a block diagram which shows the structure of a flow line simulation part. It is a flowchart explaining an example of the process by a movement path | route production | generation part in detail. The structure of a flow line adjustment part is shown. Specific examples of main flow lines An example of the main flow line set MT is shown. An example of a set SP of main flow line components is shown. An example of NU is shown. An example of mt∈t is shown. It is a flowchart explaining the process by a flow line data conversion part. The example of the input data of a flow line data conversion part is shown. It is a figure explaining step S12. It is a figure explaining step S13. It is a figure explaining step S14. It is a figure explaining step S15. A specific example of processing for the flow line t ₃ in step S15 will be described. It is a figure explaining step S16. It is a figure explaining step S17. It is a figure explaining step S18. The customer flow line set after step S18 is shown.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Attribute information database 12 Actual measurement data acquisition part 13 Moving body generation | occurrence | production interval estimation part 14 Target area estimation part 15 Area stay time estimation part 16 Inter-area passage time estimation part 17 Movement path estimation part 18 Flow line generation condition setting part 19 Flow line simulation Unit 20 main flow line extraction unit 21 flow line adjustment unit 21a flow line data conversion unit 22 flow line data storage unit

Claims (9)

The simulation flow line set, which is a set of mobile flow lines obtained by simulation, retains the main flow line extracted from the measured flow line set, which is a set of mobile flow lines obtained by measurement. A flow line processing apparatus for adjusting a simulation flow line set,
First input means for inputting the measured flow line set;
Second input means for inputting the simulation flow line set;
First main flow line extracting means for extracting a first main flow line from the measured flow line set according to a pre-given main flow line content rate;
Second main flow line extracting means for extracting a second main flow line from the simulation flow line set according to the main flow line content rate;
When the set of the second main flow lines does not match the set of the first main flow lines, the elements not included in each of the first main flow lines are determined from among the elements in each of the second main flow lines. A determination means;
The number of simulation flow lines greater than the value obtained by subtracting the number of simulation flow lines corresponding to the main flow line content rate from the number of simulation flow lines included in the simulation flow line set is selected from the simulation flow line set and selected. Removing means for removing the determined element from the simulation flow line;
For each of the first main flow lines, a simulation flow line including the first main flow line is detected from the simulation flow line set that has passed through the removing unit, and the detected simulation flow lines are classified into the same group. Classification means;
Main flow line deleting means for deleting a first main flow line other than the first main flow line corresponding to the group from the simulation flow lines in the group according to the order of the groups based on a group selection criterion given in advance;
For each of the groups, in order that the number of simulation flow lines included in the group satisfy the main flow line content rate, each of the first main movements in the simulation flow line set that has passed through the main flow line removal means. Main flow line insertion means for selecting a number of simulation flow lines that are insufficient for the group from simulation flow lines that do not include any of the lines, and inserting a first main flow line corresponding to the group into the selected simulation flow line When,
Third main flow line extracting means for extracting a third main flow line from the simulation flow line set that has passed through the main flow line inserting means according to the main flow line content rate;
When a main flow line other than the first main flow line set is included in the third main flow line set, further main flow line deletion means for deleting the main flow line from the simulation flow line including the main flow line;
Output means for outputting a set of simulation flow lines through the main flow line removing means;
A flow line processing apparatus.   The flow line processing apparatus according to claim 1, wherein the removing unit randomly selects the simulation flow line.   3. The flow line processing apparatus according to claim 1, wherein the main flow line deletion unit performs processing in order from a group having a small number of the simulation flow lines as the group selection criterion given in advance. 4.   The main flow line deleting means once removes all the elements of the first main flow lines from the simulation flow line, and then inserts the corresponding first main flow line into the simulation flow line. The flow line processing apparatus according to claim 1.   5. The flow line processing apparatus according to claim 1, wherein the main flow line insertion unit performs processing in order from a group having a small number of simulation flow lines.   The main flow line inserting means removes all the elements of each first main flow line from the selected simulation flow line, and then inserts a first main flow line corresponding to the group. Item 6. The flow line processing device according to any one of Items 1 to 5.   The flow line processing apparatus according to claim 1, wherein the main flow line insertion unit selects a different simulation flow line for each of the groups. The simulation flow line set, which is a set of flow lines of the moving body obtained by simulation, holds the main flow line estimated from the measured flow line set, which is a set of flow lines of the mobile body, obtained by measurement. A method for adjusting a simulation flow line set,
A first input step of inputting the measured flow line set;
A second input step of inputting the simulation flow line set;
A first main flow line extracting step of extracting a first main flow line from the measured flow line set according to a pre-given main flow line content rate;
A second main flow line extracting step of extracting a second main flow line from the simulation flow line set according to the main flow line content rate;
When the set of the second main flow lines does not match the set of the first main flow lines, the elements not included in each of the first main flow lines are determined from among the elements in each of the second main flow lines. A decision step;
The number of simulation flow lines greater than the value obtained by subtracting the number of simulation flow lines corresponding to the main flow line content rate from the number of simulation flow lines included in the simulation flow line set is selected from the simulation flow line set and selected. A removing step of removing the determined element from the simulation flow line;
For each of the first main flow lines, a simulation flow line including the first main flow line is detected from the simulation flow line set after the removal step, and the detected simulation flow lines are classified into the same group. A classification step;
A main flow line removing step of deleting a first main flow line other than the first main flow line corresponding to the group from a simulation flow line in the group according to the order of the groups based on a group selection criterion given in advance;
For each of the groups, in order for the number of simulation flow lines included in the group to satisfy the main flow line content rate, each of the first main movements in the simulation flow line set after the main flow line removal step. A main flow line insertion step of selecting a number of simulation flow lines that are insufficient in the group from simulation flow lines that do not include any of the lines, and inserting a first main flow line corresponding to the group into the selected simulation flow line When,
A further main flow line extraction step of extracting a third main flow line from the simulation flow line set after the main flow line insertion step according to the main flow line content rate;
When a main flow line other than the first main flow line set is included in the third main flow line set, a main flow line deletion step of deleting the main flow line from the simulation flow line including the main flow line;
An output step of outputting a set of simulation flow lines after the main flow line deletion step;
With a method. The simulation flow line set, which is a set of flow lines of the moving body obtained by simulation, holds the main flow line estimated from the measured flow line set, which is a set of flow lines of the mobile body, obtained by measurement. A program for adjusting a simulation flow line set,
A first input step of inputting the measured flow line set;
A second input step of inputting the simulation flow line set;
A first main flow line extracting step of extracting a first main flow line from the measured flow line set according to a pre-given main flow line content rate;
A second main flow line extracting step of extracting a second main flow line from the simulation flow line set according to the main flow line content rate;
When the set of the second main flow lines does not match the set of the first main flow lines, the elements not included in each of the first main flow lines are determined from among the elements in each of the second main flow lines. A decision step;
The number of simulation flow lines greater than the value obtained by subtracting the number of simulation flow lines corresponding to the main flow line content rate from the number of simulation flow lines included in the simulation flow line set is selected from the simulation flow line set and selected. A removing step of removing the determined element from the simulation flow line;
For each of the first main flow lines, a simulation flow line including the first main flow line is detected from the simulation flow line set after the removal step, and the detected simulation flow lines are classified into the same group. A classification step;
A main flow line removing step of deleting a first main flow line other than the first main flow line corresponding to the group from a simulation flow line in the group according to the order of the groups based on a group selection criterion given in advance;
For each of the groups, in order for the number of simulation flow lines included in the group to satisfy the main flow line content rate, each of the first main movements in the simulation flow line set after the main flow line removal step. A main flow line insertion step of selecting a number of simulation flow lines that are insufficient in the group from simulation flow lines that do not include any of the lines, and inserting a first main flow line corresponding to the group into the selected simulation flow line When,
A further main flow line extraction step of extracting a third main flow line from the simulation flow line set after the main flow line insertion step according to the main flow line content rate;
When a main flow line other than the first main flow line set is included in the third main flow line set, a main flow line deletion step of deleting the main flow line from the simulation flow line including the main flow line;
An output step of outputting a set of simulation flow lines after the main flow line deletion step;
A program that causes a computer to execute.

Images