KR101721622B1 - Apparatus and method for allocating security service resource - Google Patents

Abstract

The guard service resource allocation apparatus calculates an expected damage amount of at least one candidate service point where an alarm signal is generated among a plurality of registered service points, and calculates an expected damage amount of at least one candidate service point, Calculates an expected profit obtained by assigning each of the outgoing vehicles to each of the candidate service points, and allocates a plurality of outgoing vehicles to the at least one candidate service point based on the calculated expected profit.

Description

[0001] APPARATUS AND METHOD FOR ALLOCATING SECURITY SERVICE RESOURCE [0002]

The present invention relates to an apparatus and method for allocating a guard service resource.

The problem of making decisions based on uncertain information is very common. In particular, the physical security service based on the control function is based on the information generated by the intrusion detector or the fire detector installed at the customer's site, so as to guess the actual situation, dispatch the dispatcher to the site as needed, It is possible to provide a service in a corresponding manner.

Detectors have finite accuracy due to the nature of the physical sensing method, and when the property damage caused by the intrusion is caused by the security company's responsibility, there may be business loss due to the compensation obligation. In addition, the maintenance and operation of dispatch personnel are very costly, so it is necessary to minimize the size of human and material resources and to optimize the operation method when the same level of service can be maintained. Also, depending on the operational method that is dependent on subjective judgment lacking consistency, it may show a large deviation of control decisions and thereby lead to operational inefficiency.

Therefore, it is possible to minimize the operating loss of the security company by minimizing the loss of the operating company by counteracting and suppressing the actual occurrence crime by determining whether dispatching vehicle is dispatched under the uncertain information condition of the detector installed at a large number of customers, We need a way to maximize it.

Japanese Patent Publication No. 5426683 (Feb. Korean Patent Laid-Open No. 10-2013-0107629 (Oct. 20, 2013) Korean Patent Publication No. 10-2010-0083646 (Jul. 22, 2010)

SUMMARY OF THE INVENTION It is an object of the present invention to provide an apparatus and method for allocating a security service resource that minimizes operating losses of a security company and maximizes economic benefits.

According to an embodiment of the present invention, a guard service resource allocation method in a guard service resource allocation apparatus is provided. The guard service resource allocation method includes: calculating an expected damage amount of at least one candidate service point where an alarm signal is generated among a plurality of registered service points; using an expected damage amount of the at least one candidate service point Calculating an expected profit obtained by assigning each of a plurality of outgoing vehicles to each of the candidate service points, and allocating the plurality of outgoing vehicles to the at least one candidate service point based on the calculated expected profit do.

Wherein the step of allocating includes connecting an expected profit obtained by assigning each of the plurality of outgoing vehicles to each of the candidate service points to an element of a column representing a candidate vehicle service point and a row representing the outgoing vehicle to generate an expected profit matrix, Calculating a solution that maximizes the sum by selecting an expected profit exclusively for each row and each column of the expected profit matrix and assigning the plurality of dispatch vehicles to the at least one candidate service point according to the solution; .

The step of calculating the expected damage amount includes the steps of calculating a real alarm probability estimation value for an alarm signal of each candidate service point based on past accumulated data, And estimating an expected damage amount of each of the candidate service points using the real alarm probability estimation value for the alarm signal of the candidate service point.

The step of calculating the real alarm probability estimation value may include calculating the preliminary context relevance index for each of the candidate service points using the correlation index of the predictive variables for each situation, Calculating a probability of occurrence of the alarm signal according to the situation using the state transition probability model for the alarm signal and for each alarm signal of each of the candidate service points, And estimating a real alarm probability using the alarm signal occurrence probability.

The step of estimating the real alarm probability comprises the steps of calculating a probability index for each of the alarm signal of each of the candidate service points by using the context prior index and the occurrence probability of the alarm signal according to the situation, And detecting a situation corresponding to the highest probability index and the highest probability index among the context probability indexes.

The context-based state transition probability model may be generated based on the past accumulated data, and the relevance index of the context-dependent prediction variables may be calculated based on the past accumulated data and stored in the lookup table.

The situation may include a real alarm condition, a malfunction condition of the sensor that generated the alarm signal, and a false alarm condition due to other factors.

The context-based state transition probability model may be periodically verified and updated based on the verification result.

Wherein the guard service resource allocation method comprises the step of, after the step of allocating, estimating a real alarm probability estimation value for an alarm signal of a service point other than a service point where the alarm signal is determined as a real alarm among the at least one candidate service point, Changing an expected profit obtained by assigning each of the plurality of outgoing vehicles to each of the candidate service points based on a real alarm probability estimation value for an alarm signal of the remaining service point that has been changed, .

Wherein the real alarm probability estimate value is greater than 0 and less than 1, the uncertainty weight function is determined by a specific power of the real alarm probability estimate, and the particular power may be a power of an integer greater than one .

Wherein calculating the expected profit comprises calculating an arrival time of the plurality of dispatch vehicles for each of the candidate service points based on the position of the candidate service point and the position of the plurality of dispatch vehicles, Calculating an expected amount of each of the candidate service points for the plurality of dispatched vehicles using the arrival time of the plurality of dispatched vehicles for each of the plurality of dispatched vehicles and the expected damage amount of the candidate service point .

According to another embodiment of the present invention, there is provided a guard service resource allocating apparatus for allocating a moving vehicle to a plurality of service points. The guard service resource allocation apparatus includes an expected damage calculation unit, a expected profit calculation unit, and a resource allocation unit. The expected damage amount calculation unit may calculate a real alarm probability estimation value for the alarm signal of at least one candidate service point where an alarm signal is generated among the plurality of service points and calculate a real alarm probability value for the alarm signal of the at least one candidate service point And calculates an expected damage amount of the at least one candidate service point using the alarm probability estimate. The expected profit calculation unit calculates an expected profit obtained by allocating each of the plurality of dispatch vehicles to each of the at least one candidate service point using the expected damage amount of the at least one candidate service point. And the resource allocation unit allocates the plurality of dispatch vehicles to the at least one candidate service point using an expected profit obtained by allocating each of the plurality of dispatch vehicles to each of the at least one candidate service point.

The resource allocator selects an expected profit exclusively in each row and each column of the expected profit matrix having the calculated expected profit as an element and identifying the moving vehicle as a row and identifying the at least one service point as a column, And assign the plurality of dispatch vehicles to the at least one candidate service point based on the element values of the rows and columns that maximize the sum.

Wherein the expected damage amount calculation unit calculates an expected damage amount of the at least one candidate service point by using a maximum expected damage amount of the at least one candidate service point and a real alarm probability estimation value for an alarm signal of the at least one candidate service point Can be calculated.

The security service resource allocation apparatus may further include a registration unit configured to set the maximum expected damage amount based on the service items of the plurality of service points and to store the information of the plurality of service points and the set maximum expected damage amount .

The expectation damage calculation unit calculates a situation likelihood index by using a relevance index and a state transition probability model of each situation prediction parameter calculated on the basis of past accumulated data for an alarm signal of each candidate service point, The real alarm probability estimation value can be calculated by detecting a situation corresponding to the highest probability index and the highest probability index among the situation probability indexes.

The expectation damage calculation unit calculates the preliminary context relevance index by using the relevance index of the contextual predictive variables and calculates an occurrence probability of the alert signal according to the context using the contextual context transition probability model, The likelihood index according to the situation can be calculated using the pre-context relevance index according to the situation and the occurrence probability of the alarm signal according to the situation.

Wherein the resource allocation unit changes the real alarm probability estimation value for the alarm signal of the service point except for the service point where the alarm signal is determined as the real alarm among the at least one candidate service point, according to the uncertainty weight function, And to change the expected profit obtained by assigning each of the vehicles to each of the at least one candidate service point.

Wherein the real alarm probability estimate value is greater than 0 and less than 1, the uncertainty weight function is determined by a specific power of the real alarm probability estimate, and the particular power may be a power of an integer greater than one .

Wherein the expected profit calculation unit calculates the arrival time of the plurality of dispatch vehicles for each of the candidate service points based on the position of the candidate service point and the position of the plurality of dispatch vehicles, The expected amount of each of the candidate service points for the plurality of dispatched vehicles can be calculated using the arrival time of the dispatched vehicle and the expected damage amount of the candidate service point.

According to the embodiment of the present invention, it is possible to obtain a probability distribution of elements having uncertainty based on the log data accumulated in the operation process with respect to the uncertainty in the control operation, and calculate the expected damage amount And optimizing the allocation of the dispatched vehicle based on the expected profit, thereby optimizing the operational efficiency.

1 is a diagram schematically illustrating an apparatus for allocating a guard service resource according to an embodiment of the present invention.
FIG. 2 illustrates a detailed configuration of a guard service resource allocation apparatus according to an embodiment of the present invention. Referring to FIG.
3 is a flowchart illustrating a method of allocating a guard service resource according to an embodiment of the present invention.
4 is a diagram illustrating an example of a probability distribution of a time required for a crime according to an embodiment of the present invention.
5 is a diagram illustrating an example of a weight function for uncertainty according to an embodiment of the present invention.
6 is a flowchart illustrating a method of calculating a real alarm probability estimate value according to an embodiment of the present invention.
FIG. 7 is an example of a preliminary distribution chart showing the occurrence frequency of a specific situation according to an embodiment of the present invention.
8 is a diagram illustrating an example of a transition time distribution according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification and claims, when a section is referred to as "including " an element, it is understood that it does not exclude other elements, but may include other elements, unless specifically stated otherwise.

Now, an apparatus and method for allocating a guard service resource according to an embodiment of the present invention will be described in detail with reference to the drawings.

1 is a diagram schematically illustrating an apparatus for allocating a guard service resource according to an embodiment of the present invention.

Referring to FIG. 1, a guard service resource allocation apparatus 100 provides a security expense and a dispatch service of a service point 200 registered according to a contract with a customer.

Each service point 200 is provided with at least one detector 210 for security expense and dispatch service. The at least one detector 210 may include, for example, a detector capable of detecting an intruder, a detector capable of detecting a fire, a detector capable of detecting a natural disaster such as a typhoon, a flood, and an earthquake.

The guard service resource allocation apparatus 100 allocates a dispatch vehicle assignment (ID) to a service point 200 where an alarm signal is generated from the sensor 210 among the registered service points 200, based on data accumulated through the security expense and dispatch service And the optimal service point assignment of the dispatch vehicle.

FIG. 2 is a detailed configuration diagram of a guard service resource allocation apparatus according to an embodiment of the present invention, and FIG. 3 is a flowchart illustrating a guard service resource allocation method according to an embodiment of the present invention.

2, the guard service resource allocation apparatus 100 includes a registration unit 110, a storage unit 120, an expected damage calculation unit 130, an expected profit calculation unit 140, and a resource allocation unit 150 .

The registration unit 110 registers the information of the contracted service point. Based on the service item of the contracted service point, the registration unit 110 sets a maximum expected damage amount that can be generated when an accident such as an intrusion, a fire, a natural disaster, or the like occurs and stores the information of the contracted service point, (120). The information of the contracted service point may include operator information, the location of the service point, service items, business hours, and the like.

3, when the alarm signal is generated from the detector 210 (S310), the expected damage value calculation unit 130 calculates a maximum damage value of the service point 200 at which the alarm signal is generated, (S320). For example, if an alarm signal is generated at S service points, the maximum expected damage amount list includes the maximum damage amount per S service points.

The expected damage amount calculation unit 130 adds the maximum expected damage amount of the service point corresponding to the newly generated alarm signal to the maximum expected damage amount list or the maximum expected damage amount of the service point where the false alarm occurs if the alarm signal is determined to be false You can delete the amount from the maximum expected damage amount list.

The expected damage calculation unit 130 calculates a probability that the alarm signal of the service point corresponds to the real alarm, that is, a real alarm probability estimate value for the alarm signal of the service point (S330). The alarm signal has a finite accuracy and may or may not be visible. The expected damage calculation unit 130 may calculate a real alarm probability estimation value for the alarm signal of the service point based on the data accumulated through the security expense and the dispatch service.

Next, the expected damage calculation unit 130 calculates the expected damage amount of the service point where the alarm signal is generated by using the real alarm probability estimation value for the alarm signal of each service point and the maximum expected damage amount of the service point (S340 ). For example, the expected damage amount of each service point can be calculated as shown in Equation (1).

In Equation (1), L_exp, i represents the expected damage amount of the service point s_i, L_i represents the maximum expected damage amount of the service point s_i, and w_i represents a real alarm probability estimate value for the alarm signal of the service point s_i.

The expected damage calculation unit 130 generates an expected damage amount vector having an expected damage amount of each service point where an alarm signal is generated as an element (S350).

The expected profit calculation unit 140 calculates an expected profit that may occur when the operated moving vehicle is assigned to a service point where an alarm signal is generated (S360).

For example, the time required for the attack from the detector 210 to the point of time of the intruder escape after the alarm signal is firstly follows the probability distribution of the estimated time of the crime estimated based on the operation log data.

4 is a diagram illustrating an example of a probability distribution of a time required for a crime according to an embodiment of the present invention.

The probability distribution D (t) of the time required for such a crime can be expressed by Equation (2).

It is assumed that if the guard arrives before the end of the crime, based on the alert signal time of the service point, the amount of damage can be prevented or minimized to a negligible level.

The expected profit when the dispatched vehicle arrives at the service point before the time required for the crime after the first alarm signal is generated may correspond to the negative value of the expected damage amount as shown in Equation (3).

In Equation (3), P_exp, i represents the expected profit when the dispatching vehicle arrives at the service point s_i before the time required for the commutation after the first alarm signal is generated.

The arrival time of the dispatching vehicle for the departing vehicle to depart from the position at the time of generation of the alarm signal of the service point and to reach the service point occurs.

The expected profit calculation unit 140 calculates the expected arrival time of the dispatched vehicle based on the position of the dispatched vehicle at the time of generating the alarm signal of the service point and the position of the service point where the alarm signal is generated, The profit can be calculated as shown in Equation (4).

In Equation (4), P_m, i (t_t) represents the expected profit obtained by assigning the dispatching vehicle c, m to the service point s_i. t_t represents the arrival time.

In other words, the expected profit obtained by assigning the dispatch vehicle to the service point can be expressed as the product of the anticipated profit when the dispatched vehicle arrives at the service point before the commissioning time after the first alarm signal, .

The expected profit calculation unit 140 calculates an expected profit vector composed of the expected profit obtained by assigning the dispatch vehicle to each service point based on Equation (4).

The expected profit calculation unit 140 generates an expected profit matrix from the calculated expected profit vector for all dispatch vehicles (S370). The dispatch vehicle and the service point are variables that identify rows and columns in the expected profit matrix, and the values retrieved by rows and columns represent the expected profit obtained when the dispatch vehicle assigns to the service point. For example, when there are M service vehicles and S service points where alarm signals are generated, the expected profit matrix P can be expressed by Equation (5).

In Equation (5), P_j, i represents the expected profit when the dispatch vehicle c, j is assigned to the service point s_i.

The resource allocation unit 150 determines an optimal service point to which each moving vehicle is allocated by using the expected profit matrix generated by the expected profit calculation unit 140 (S380). The resource allocation unit 150 selects a single expected profit exclusively in each row and each column of the expected profit matrix and obtains a solution for maximizing the sum and assigns each dispatch vehicle to the service point according to the obtained solution. The resource allocating unit 150 can obtain a solution for maximizing the sum of the expected profit for each row and each column exclusively by using the Munkres method.

For example, suppose that the expected profit matrix is equal to Equation (6). (1, 3), (2, 2), (3, 1), (4), and (5) 1)].

The resource allocating unit 150 allocates the moving vehicle c, 1 to the service point s_3 based on the solutions [(1,3), (2, 2), (3, 1), Assigning the dispatch vehicle c, 2 to the service point s_2, allocating the dispatch vehicle c, 3 to the service point s_1, and assigning the dispatch vehicle c, 4 to the service point s_4.

The dispatching agent moves to the assigned service point using the dispatch vehicle. It is confirmed by the dispatching personnel that the alarm signal of the corresponding service point is a real alarm.

The resource allocation unit 150 assigns a weight to the uncertainty in the real alarm probability estimate value for the alarm signal for the service points other than the service point where the alarm signal is determined as the real alarm from the dispatching agent for each dispatching vehicle, The expected profit matrix is changed as shown in Equation 8 (S390).

In Equation (8), p_i, j denotes an element located in the i-th row and j-th column of the expected profit matrix shown in Equation (5), and w (p_i, j) denotes an expected profit given according to a weight function for uncertainty Represents the amount of change.

That is, the expected profit for the service points other than the service point where the alarm signal is determined as the real alarm is changed based on the equation (8), and the expected profit matrix is changed eventually.

5 is a diagram illustrating an example of a weight function for uncertainty according to an embodiment of the present invention.

For example, the weight function for the uncertainty can be given as shown in FIG. 5, and according to the weight function for the uncertainty, the alarm signal for the remaining service points except for the service point, The alarm probability estimate value is changed.

When the weight function for the uncertainty is given as shown in FIG. 5, the real alarm probability estimate value for the alarm signal of any one of the service points other than the service point where the alarm signal is determined as the real alarm is 0.8 The estimated value of the real alarm probability for the alarm signal of the service point is changed to 0.4.

In this way, the expected profit matrix is changed while changing the real alarm probability estimation value for the alarm signal of the service points other than the service point where the alarm signal is determined as the real alarm.

In this manner, the resource allocation unit 150 lowers the real alarm probability estimation value for the alarm signal of the remaining service points except for the service point where the alarm signal is determined as the real alarm. That is, since the dispatch vehicle (service dispatcher) is already assigned to the corresponding service point, another dispatch vehicle (service dispatcher) need not be allocated. Therefore, by changing the expected profit matrix, the alarm signal lowers the real alarm probability estimate value for the alarm signal of the service points other than the service point where the alarm signal is determined as the real alarm, so that the other service vehicle is not allocated to the corresponding service point.

6 is a flowchart illustrating a method of calculating a real alarm probability estimate value according to an embodiment of the present invention.

6, when the alarm signal is generated from a service point, the expected damage value calculation unit 130 calculates an expected damage value based on a pre-knowledge-based specific situation according to a metadata prediction parameter that defines an event currently occurring according to an alarm signal. A pre-context association index indicating the degree of association is calculated (S610). The metadata prediction parameters may include, for example, the time of occurrence, the location of occurrence, the frequency of occurrence, and the like.

The data of the event in which the alarm signal is generated may include the first generated alarm signal, the subsequent alarm signal, the generated point metadata, and the type of the recorded situation, and the data of the event Can be accumulated and a large data set can be obtained.

The data set can be classified into a detailed set according to the situation category. For example, a large set of data can be categorized into real alarm conditions, detector malfunction situations, and false alarms caused by similar intrusions (eg, small animals).

The temporal, spatial, and social conditions of a situation occurrence point may have a relatively high or low correlation with the type of specific situation. For example, there is a high probability that an intruder will be stolen in the middle of the day, and that certain industries, such as precious metals and mobile phone stores, are likely to be targets of intrusion.

FIG. 7 is an example of a preliminary distribution chart showing the occurrence frequency of a specific situation according to an embodiment of the present invention.

Based on the preliminary distribution of the occurrence probability [F _X (x)] of the occurrence frequency of the specific situation according to the time (x) shown in FIG. 7, the association index between the time (x) and the specific situation is calculated and stored in the lookup table .

In this way, a lookup table storing a predictive variable and a correlation index of a specific situation is generated as a prior knowledge base of past occurrence data according to a situation, and a correlation index with each situation can be determined through a lookup table.

On the other hand, if the proportion of the situation occurrence cases of the specific variable-specific category values is significantly different from that of the other situation occurrence cases, the relevance index corresponding to the specific variables is determined through the siginificance test, Can be stored in a table.

The expected damage amount calculation unit 130 may calculate a pre-condition relevancy index for the alarm signal for each situation from the sum of the context relevance indexes of the metadata prediction parameters that define the current occurrence event as shown in Equation (9).

In Equation (9), A_p_i and S_j represent the state softness index with respect to the state S_j for the predictive variable p_i.

Referring to FIG. 6 again, the expected damage amount calculation unit 130 calculates the occurrence probability of the alarm signal currently generated using the state transition probability model (S620).

The occurrence of an alarm signal according to the past situation can be regarded as an instance of a probabilistic representation of a situation-specific generator model.

In the past, the alarm signal is used as an input of the Markov model for modeling the state transition probability. The signal stream of the alarm signal is generated from the signal vector sequence from the point of time when the first alarm signal is generated to the end of the situation, Transition time can be characterized.

8 is a diagram illustrating an example of a transition time distribution according to an embodiment of the present invention. 8 (a), 8 (b) and 8 (c), the x-axis represents time and the y-axis represents probability (%).

As shown in FIGS. 8A, 8B and 8C, based on the data input of the events constituted by the transition target state and the transition time in each state, the transition probability between states and the transition time distribution Can be obtained, and a situation-specific state transition model can be generated.

The generated state transition model is verified through simulation and can be updated due to changes in the nature of the situation and structure due to social and natural factors. After generating the random sample data set through the simulation process, the distance between the state vector sequence probability distribution of the random sample data set and the state vector sequence probability distribution of the actual situation data set is calculated, The state transition model can be verified. Here, the state vector sequence probability distribution means the probability distribution of the state vector sequence. The state vector sequence means a sequence of alarm signals, and one alarm signal corresponds to one element of the state vector. If the calculated distance value is greater than the set threshold value, the state transition model of the corresponding state can be updated based on the latest data.

The expected damage calculation unit 130 may calculate the occurrence probability of the alarm signal currently generated under the Hypothesis of the specific state transition model among the state transition models as shown in Equation (10).

In Equation (10), M_S_j denotes a specific situation model, and Sig_k denotes an alarm signal currently generated. s_0 represents an initial state, t_0 → 1 represents a transition time from s_0 to s_1, s_1 and s_2 represent transition states, and t_1 → 2 represents a transition time from s_1 to s_2.

That is, the occurrence probability of the alarm signal currently generated in the specific situation model can be calculated as the initial state probability, the transition time probability, and the product of the transition object state probability.

The expected damage calculation unit 130 calculates the occurrence probability of the alarm signal currently generated under the hypothesis of the state transition model.

The expectation damage calculation unit 130 calculates a likelihood score for the alert signal based on the calculated context-related prior art context index and the occurrence probability of the alert signal as shown in Equation 11 (S630).

In Equation (11), A_S_j represents a pre-context correlation index for a situation S_j, and P_S_j represents a probability of occurrence of an alarm signal (e) for a situation S_j. W_A and w_P respectively represent weights of the pre-context indexes and weights of occurrence probabilities, w_A and w_P may be predetermined values, and may be arbitrarily changed.

The expected damage amount calculation unit 130 extracts the highest possibility index among the computed situation possibility indexes (S640), and detects a situation corresponding to the highest possibility index (S650).

The expected damage calculation unit 130 calculates a real alarm probability estimation value based on the highest probability index and the corresponding situation (S660). For example, when the highest likelihood index and the corresponding situation are detected, the expected damage calculation unit 130 calculates a real alarm probability estimate to have a value between 0 and 1 through appropriate standardization for the highest probability index Can be calculated.

At least some functions of the apparatus and method for allocating cost resources according to the above-described embodiments of the present invention may be realized by hardware or software combined with hardware. For example, a processor implemented as a central processing unit (CPU) or other chipset, microprocessor, etc. may be used as a registration unit 110, an expected damage calculation unit 130, an expected profit calculation unit 140, And a memory or a storage device may perform the function of the storage unit 120. [0034] And a transceiver coupled to the processor may receive an alarm signal from the detector.

The embodiments of the present invention are not limited to the above-described apparatuses and / or methods, but may be implemented through a program for realizing functions corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded, Such an embodiment can be readily implemented by those skilled in the art from the description of the embodiments described above.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, It belongs to the scope of right.

Claims (20)

A guard service resource allocation method in a guard service resource allocation apparatus,
Calculating a real alarm probability estimation value for each alarm signal of each of at least one candidate service point where an alarm signal is generated among a plurality of registered service points based on past accumulated data,
Calculating an expected damage amount of each of the candidate service points using a maximum expected damage amount of each of the candidate service points and a real alarm probability estimation value for an alarm signal of each of the candidate service points,
Calculating an expected profit obtained by allocating each of a plurality of running vehicles that are operating to each of the candidate service points using an expected damage amount of the at least one candidate service point,
Allocating at least one of the plurality of moving vehicles to the at least one candidate service point based on the calculated expected profit,
Wherein the real alarm probability estimation value for the alarm signal of the service point other than the service point where the alarm signal is determined as the real alarm among the at least one candidate service point to which the at least one vehicle is assigned is lowered according to the uncertainty weight function Step, and
And changing the expected profit obtained by assigning each of the plurality of outgoing vehicles to each of the candidate service points based on the estimated real alarm probability value of the alarm signal of the remaining service point, Steps to lower the probability of assigning
And allocating a guard service resource. The method of claim 1,
The allocating step
Generating an expected profit matrix by connecting an expected profit obtained by assigning each of the plurality of outgoing vehicles to each of the candidate service points to an element of a column representing a candidate vehicle service point and a row representing a departure vehicle,
Selecting a single expected profit exclusively in each row and each column of the expected profit matrix and calculating a solution for maximizing the sum; and
And allocating the at least one mobile vehicle to the at least one candidate service point according to the solution. delete The method of claim 1,
The step of calculating the real alarm probability estimate value
Calculating a context-dependent prior context context index using an association index of context-dependent prediction parameters for an alert signal of each of the candidate service points;
Calculating a probability of occurrence of the alarm signal for each situation using the state transition probability model for each alarm signal of each of the candidate service points; and
And estimating a real alarm probability using an occurrence probability of the alarm signal according to the situation and a pre-context correlation index for each of the alarm signals at each of the candidate service points. 5. The method of claim 4,
The step of estimating the real alarm probability
Calculating a likelihood index for each of the candidate service points using the probability of occurrence of the alarm signal according to the situation and the pre-context correlation index for each situation with respect to the alarm signal of each of the candidate service points; and
And detecting a situation corresponding to the highest probability index and the highest probability index among the context probability indexes. 5. The method of claim 4,
Wherein the context-based state transition probability model is generated based on the past accumulated data, and the relevancy index of the context-dependent prediction variables is calculated based on the past accumulated data and stored in a look-up table. 5. The method of claim 4,
Wherein the situation includes a false alarm situation, a malfunction condition of the detector that generated the alarm signal, and a false alarm condition due to other factors. 5. The method of claim 4,
Wherein the state transition probability model is periodically verified and updated based on the verification result. delete The method of claim 1,
Wherein the real alarm probability estimate value is greater than 0 and less than 1, the uncertainty weight function is determined by a specific power of the real alarm probability estimate, and the specific power is a power of an integer greater than one, Resource allocation method. The method of claim 1,
The step of calculating the expected profit
Calculating an arrival time of each of the plurality of dispatch vehicles for each of the candidate service points, based on the position of the candidate service point and the position of the plurality of dispatch vehicles, and
Calculating an expected amount of each of the candidate service points with respect to the plurality of dispatched vehicles using the arrival time of the plurality of dispatched vehicles for each of the candidate service points and the expected damage amount of the candidate service point A method of allocating a guard service resource. 1. A guard service resource allocation apparatus for allocating a departure vehicle to a plurality of service points,
Calculating a real alarm probability estimation value for the alarm signal of at least one candidate service point where an alarm signal is generated among the plurality of service points based on past accumulated data, An expected damage amount calculating unit for calculating an expected damage amount of the at least one candidate service point by using a real alarm probability estimation value for an alarm signal of a candidate service point,
An expected profit calculation unit for calculating an expected profit obtained by allocating each of the plurality of outgoing vehicles to each of the at least one candidate service point using the expected damage amount of the at least one candidate service point,
Allocating at least one outgoing vehicle among the plurality of outgoing vehicles to the at least one candidate service point by using an expected profit obtained by allocating each of the plurality of outgoing vehicles to each of the at least one candidate service point,
/ RTI >
Wherein the resource allocation unit estimates a real alarm probability estimation value for an alarm signal of a service point other than the service point where the alarm signal is determined as a real alarm among the at least one candidate service point to which the at least one moving vehicle is allocated as an uncertainty weight And changing an expected profit obtained by assigning each of the plurality of outgoing vehicles to each of the candidate service points based on a real alarm probability estimation value for the alarm signal of the remaining service points, And reduces the probability of allocating a plurality of dispatch vehicles. The method of claim 12,
The resource allocator selects an expected profit exclusively in each row and each column of the expected profit matrix having the calculated expected profit as an element and identifying the moving vehicle as a row and identifying the at least one service point as a column, And assigns the at least one moving vehicle to the at least one candidate service point based on element values of rows and columns that maximize the sum. delete The method of claim 12,
Setting a maximum expected damage amount based on service items of the plurality of service points, and storing information of the plurality of service points and the set maximum expected damage amount,
Further comprising a resource allocation unit for allocating resource allocation to the resource. The method of claim 12,
The expectation damage calculation unit calculates a situation likelihood index by using a relevance index and a state transition probability model of each situation prediction parameter calculated on the basis of past accumulated data for an alarm signal of each candidate service point, And calculates the real alarm probability estimation value by detecting a situation corresponding to the highest probability index and the highest probability index among the context probability indexes. 17. The method of claim 16,
The expectation damage calculation unit calculates the preliminary context relevance index by using the relevance index of the contextual predictive variables and calculates an occurrence probability of the alert signal according to the context using the contextual context transition probability model, And calculates the context-dependent likelihood index using the context-related context-relatedness index and the probability of occurrence of the alert signal according to the context. delete The method of claim 12,
Wherein the real alarm probability estimate value is greater than 0 and less than 1, the uncertainty weight function is determined by a specific power of the real alarm probability estimate, and the specific power is a power of an integer greater than one, Resource allocation device. The method of claim 12,
Wherein the expected profit calculation unit calculates the arrival time of the plurality of dispatch vehicles for each of the candidate service points based on the position of the candidate service point and the position of the plurality of dispatch vehicles, And calculates an expected amount of each of the candidate service points with respect to the plurality of dispatched vehicles by using an arrival time of the dispatched vehicle and an expected damage amount of the candidate service point.

Images