JP5013506B2 - Application program execution method, apparatus and program - Google Patents

Description

  The present invention particularly relates to a technique for executing an application program using constraint programming (hereinafter simply referred to as an application).

  Constraint programming is a framework of methods and ideas for solving problems that find and find answers that satisfy constraints. In constraint programming, two steps are taken: declaration of a problem to be solved and search for a solution that satisfies the problem. Here, the declaration of a problem to be solved means that a variable called a “constraint variable” consisting of a set of possible values is defined and a “relational expression between constraint variables” representing a constraint condition is defined. Further, a general-purpose program called “constraint satisfaction solver” is generally used for searching for solutions.

  In the constraint satisfaction solver, it is possible to efficiently search for solutions that satisfy the constraint conditions using various strategies by converting constraint variables and constraint conditions into the internal representation specific to the constraint satisfaction solver through the interface. It has the characteristic of being.

There are a plurality of constraint satisfaction solvers, such as those that can be used for free and those that can be used for commercial purposes. Each constraint satisfaction solver has a difference in development environment, execution environment, interface, etc., but the basic processing flow is the same. Many constraint satisfaction solvers generate constraint initialization solvers, initialize them, define constraint variables to convert constraint variables to an internal representation, constraint expressions that define constraint conditions, and add constraint conditions to constraint satisfaction solvers to enable constraints The steps of constraint propagation to obtain a solution and obtain a solution set satisfying the constraint are taken.
"Introduction to Constraint Programming", [online], NTT DATA Sekisui Systems, Inc. [Search May 29, 2006], Internet <URL: http://www.isac.co.jp/iz/ tutorial.html>

  An application using constraint programming often incorporates, as an internal structure, a declaration of a problem to be solved consisting of a constraint variable definition and a constraint condition definition by an internal representation specific to the constraint satisfaction solver used. In such a case, if a change that changes the constraint condition assumed at the time of development occurs, modification of the application is required. However, the cost is generally high and it is not easy.

In the present invention, for solving the above problems, the constraint definition storage unit for storing a constraint definition that application handles, and a data domain storing unit for storing with constraint variable name data domain that the application is handled provided, when the application execution The constraint definition is read from the constraint definition storage unit, the constraint variable that matches the condition of the predetermined constraint variable name is extracted from the constraint definition, and the constraint variable name of the extracted constraint variable is extracted from the data domain storage unit as a key. Obtain the data domain of the variable , convert the constraint variable in the obtained data domain into an internal representation specific to the constraint satisfaction solver, store this along with the constraint variable name, extract the relational expression from the constraint definition, Using the constraint variables converted to the internal representation specific to the stored constraint satisfaction solver, the extracted function By constructing an expression as constraint satisfaction solver specific internal representation, the application itself, the declaration defining a problem to be solved consists of the definition and constraints of the constraint variables were not necessary to incorporate as an internal structure.

  According to the present invention, the declaration of the problem to be solved consisting of the definition of the constraint variable and the definition of the constraint condition can be dynamically acquired and constructed at the time of application execution. When the required constraints change, it is possible to deal with the above-mentioned changes by rewriting only the constraints defined independently of the application, and the effect of reducing the cost of modifying the application source code Can be expected. Along with this, it can be expected to reduce the cost of recompiling and reinstalling the source code. However, if there are changes that cannot be dealt with only by changing the constraint definition, such as addition, deletion, or change of data items, it may be necessary to modify the source code.

  FIG. 1 shows an example of an embodiment of an application (program) execution apparatus according to the present invention, here an example realized on a known computer. In the figure, 1 is a computer, 2 is an input device, and 3 is a display. Device.

  The computer 1 provides an execution environment for a predetermined application (program) 4 using constraint programming and a predetermined constraint satisfaction solver 5 used by the application 4, and is installed via a recording medium or a communication line. 2 and the program corresponding to the flowchart shown in FIG. 3. At this time, the constraint definition storage unit 6, the data domain storage unit 7, and the constraint processing function unit 8 are configured.

  The input device 2 includes a keyboard, a mouse, and the like, and is used by an operator (inspector) to input predetermined information such as data to be stored in the constraint definition storage unit 6 or the data domain storage unit 7 in the computer 1. The display device 3 includes an image display device such as a liquid crystal display or a CRT, and is used for presenting a solution satisfying a constraint condition that is a predetermined processing result in the computer 1, for example, a processing result by the application 4, to the operator.

  The constraint definition storage unit 6 is a block between constraint variables handled by the application 4 that is directly input from the input device 2 or read from a storage medium (not shown) or input from another device or the like via a communication medium (not shown). A constraint definition consisting of a set of constraint conditions represented by the relational expression is stored. Further, the data domain storage unit 7 stores a data domain, which is a set of values that can be taken by the constraint variable, input in the same manner as described above, together with the constraint variable name that is the name of the constraint variable.

  The constraint processing function unit 8 inputs information on the location of each constraint definition and data domain from the application 4, and based on the input information, the constraint definition storage unit 6 and the data domain storage unit 7 respectively input the constraint definition and data domain. Load a domain. Then, the constraint satisfaction solver 5 converts the read constraint definition and data domain into an internal representation specific to the constraint satisfaction solver 5 and outputs the internal representation of the obtained constraint to the application 4.

  In order to realize these functions, the constraint processing function unit 8 is configured by a constraint variable management unit 81 and a constraint condition construction unit 82.

  The constraint variable management means 81 reads the constraint definition from the constraint definition storage unit 6, extracts the constraint variable from the constraint definition, and uses the constraint variable name of the constraint variable as a key from the data domain storage unit 7 to the data domain of the constraint variable Is converted into an internal representation specific to the constraint satisfaction solver 5 and stored together with the constraint variable name.

  The constraint condition construction unit 82 reads the constraint definition from the constraint variable management unit 81 or the constraint definition from the constraint definition storage unit 6, extracts a relational expression from the constraint definition, and stores it in the constraint variable management unit 81. The extracted relational expression is constructed as an internal representation specific to the constraint satisfaction solver 5 using the constraint variables converted into the internal representation specific to the constraint satisfaction solver.

  FIG. 2 is a flowchart showing the processing procedure of the constraint variable management means 81, and FIG. 3 is a flowchart showing the processing procedure of the constraint condition construction means 82. Hereinafter, the operation of the apparatus of the present invention will be described in detail.

  First, the constraint variable management unit 81 performs processing. The constraint variable management unit 81 reads the constraint definition from the constraint definition storage unit 6 (s1), and extracts lexical characters one by one from the read constraint definition (s2). If the extracted lexical phrase matches a predetermined constraint variable name condition (for example, a character string starting with an alphabet) (s3), the character string, that is, the constraint variable name is stored, and the constraint variable name is stored as a key. The data domain of the constraint variable is acquired from the data domain storage unit 7 (s4). Subsequently, the lexical word is converted into an internal representation unique to the constraint satisfaction solver 5 (s5). The constraint variable name and the converted constraint variable entity or reference are stored in association with each other (s6). This operation is performed for all the tokens (s7), and then the entity of the constraint definition or a reference to it is passed to the constraint condition construction means 82 (s8), and the process is terminated.

  Subsequently, the constraint condition construction unit 82 performs processing. The constraint condition construction means 82 reads the constraint definition (s11), and extracts relational expressions one by one from the read constraint definition (s12). Then, the extracted relational expression is parsed (s13), and is constructed as an internal representation unique to the constraint satisfaction solver 5 by a predetermined method.

  Specifically, each lexical phrase used in a relational expression is read one by one, and the read lexical expression is classified into a constraint variable name, an operator, and parentheses, and an operation is performed so as not to impair the meaning of the relational expression. Considering the priority of the child and the position of the parenthesis, by appropriately calling the interface defined in the constraint satisfaction solver 5 for the constraint between the constraint variables corresponding to the operator, the constraint is represented in the constraint satisfaction solver 5 as an internal representation. To construct. The entity or reference of the constraint variable necessary for this processing is acquired from the constraint variable management means 81 using the constraint variable name as a key. Then, the relational expression based on the internal expression unique to the constraint satisfaction solver 5 thus constructed is added to the constraint satisfaction solver 5 (s14).

  This process is performed individually for all the relational expressions (s15), and the process ends when the relational expressions disappear (s16).

  When the above processing ends, the constraint processing function unit 8 passes the internal representation of the constraint satisfaction solver 5 as an output to the application 4 and ends the processing.

  Next, details of the processing in the present invention will be described with an example.

  In this example, the constraint between data handled by the application 4 (constraint definition) is described in the file “const.txt” stored in the constraint definition storage unit 6, and the domain of each data is stored in the data domain storage unit 7. It is assumed that it is described in the file “data.txt”. The contents of “const.txt” and “data.txt” are shown in FIGS. 4 and 5, respectively.

  Hereinafter, the processing procedure until the constraint is constructed when the constraint processing function unit 8 inputs these files will be described below.

  1. The constraint variable management unit 81 of the constraint processing function unit 8 inputs the location information of the file “const.txt” and the file “data.txt” from the application 4.

  2. The constraint variable management unit 81 inputs each file from the constraint definition storage unit 6 and the data domain storage unit 7 based on the input location information.

  3. “A” is extracted as a constraint variable name.

  4). The range {1, 3, 5} is obtained by referring to the file “data.txt” using the variable name “a” as a key.

  5). “A” is converted into an internal representation unique to the constraint satisfaction solver 5 using the interface of the constraint satisfaction solver 5, and a reference to the converted entity is stored using the constraint variable name as a key.

  6). “Bb” is extracted as the constraint variable name.

  7). The range {2, 3, 4} is acquired by referring to the file “data.txt” using the variable name “bb” as a key.

  8). “Bb” is converted into an internal representation unique to the constraint satisfaction solver 5 using the interface of the constraint satisfaction solver 5, and a reference to the converted entity is stored using the constraint variable name as a key.

  9. “C1” is extracted as the constraint variable name.

  10. The range {3, 4} is acquired by referring to the file “data.txt” using the variable name “c1” as a key.

  11. “C1” is converted into an internal representation unique to the constraint satisfaction solver 5 using the interface of the constraint satisfaction solver 5, and a reference to the converted entity is stored using the constraint variable name as a key.

  12 The constraint variable management unit 81 recognizes that there is no subsequent token, passes the file “const.txt” to the constraint condition construction unit 82, and ends the process.

  13. The constraint condition construction unit 82 inputs the file “const.txt” from the constraint variable management unit 81.

  14 The relational expression “(a == 3) ‖ (bb! = 2)” is extracted.

  15. A constraint definition corresponding to the relational expression “(a == 3) ‖ (bb! = 2)” is constructed as an internal representation unique to the constraint satisfaction solver 5 using the interface of the constraint satisfaction solver 5 and output to the constraint satisfaction solver 5 To do. At this time, the lexicals “a” and “bb” are obtained with reference to the constraint variable name as a key and used for construction of the constraint definition.

  16. The relational expression “(c1> = 4) && (a <= 5)” is extracted.

  17. A constraint definition corresponding to the relational expression “(c1> = 4) && (a <= 5)” is constructed using the interface of the constraint satisfaction solver 5 and output to the constraint satisfaction solver 5. At this time, the lexical phrases “c1” and “a” are obtained with reference to the constraint variable name as a key and used to construct the constraint definition.

  18. The constraint condition construction unit 82 recognizes that there is no subsequent constraint condition, and ends the process.

  With the above procedure, it is possible to dynamically acquire and construct independently defined constraint conditions when the application 4 is executed.

  Thereafter, by appropriately calling the interface of the constraint satisfaction solver 5, it is possible to acquire a solution set that satisfies the constraints indicated in the constraint conditions.

  In the description so far, the processing procedure of the invention was shown without specifying a constraint satisfaction solver, but as shown in the previous section, there is no essential difference in the flow of processing in various constraint satisfaction solvers. Therefore, the above procedure can be executed by appropriately using the interface of each constraint satisfaction solver.

  In this example, the constraint expression method between the data in the constraint definition file and the meaning thereof are those used in the description of conditional expressions in programming languages such as C language and Java (registered trademark). Originally, this is what the constraint processing function unit 8 requires, and it is a necessary condition that the constraint processing function unit 8 can process. The same applies to the data structure of the file and database for managing the domain, and the restriction processing function unit 8 is required to be able to process.

<Other examples of embodiment>
In the above-described embodiment, the configuration in which the application and the constraint processing function unit are independent has been described. However, a configuration in which all or a part of the function of the constraint processing function unit is incorporated in the application is also possible.

The block diagram which shows an example of embodiment of the execution apparatus of the application of this invention Flow chart showing processing procedure of constraint variable management means Flow chart showing processing procedure of constraint condition construction means A figure showing an example of a constraint condition expression Figure showing an example of a data file

Explanation of symbols

  1: computer, 2: input device, 3: display device, 4: application (program), 5: constraint satisfaction solver, 6: constraint definition storage unit, 7: data domain storage unit, 8: constraint processing function unit, 81: Constraint variable management means, 82: Constraint condition construction means.

Claims (3)

A constraint satisfaction solution that satisfies a constraint definition consisting of a set of constraint conditions represented by a relational expression between constraint variables is specified from the data domain that is a set of possible values of the constraint variable. In an application program execution method obtained using a constraint satisfaction solver that searches for a constraint satisfaction solution by converting to the internal representation of
Computer
Reading the constraint definition from the constraint definition storage unit that stores the constraint definition handled by the application program ;
Extracting a constraint variable matching a condition of a predetermined constraint variable name from the constraint definition;
Acquiring the data domain of the constraint variable using the constraint variable name of the extracted constraint variable as a key from the data domain storage unit that stores the data domain handled by the application program together with the constraint variable name that is the name of the constraint variable;
Converting the obtained constraint variable in the data domain into an internal representation specific to the constraint satisfaction solver, and storing this together with the constraint variable name;
Extracting a relational expression from the constraint definition;
Using the constraint variable converted into the stored internal expression specific to the constraint satisfaction solver, and constructing the extracted relational expression as an internal expression specific to the constraint satisfaction solver. Execution method. A constraint satisfaction solution that satisfies a constraint definition consisting of a set of constraint conditions represented by a relational expression between constraint variables is specified from the data domain that is a set of possible values of the constraint variable. In an execution device of an application program obtained by using a constraint satisfaction solver that searches for a constraint satisfaction solution by converting to an internal representation of
A constraint definition storage unit that stores constraint definitions handled by the application program ;
A data domain storage unit that stores a data domain handled by the application program together with a constraint variable name that is a name of a constraint variable;
Read the constraint definition from the constraint definition storage unit, extract the constraint variable that matches the condition of the predetermined constraint variable name from the constraint definition, and use the constraint variable name of the constraint variable as a key to store data of the constraint variable from the data domain storage unit A constraint variable management means for acquiring a domain , converting the constraint variable in the acquired data domain into an internal representation specific to the constraint satisfaction solver, and storing the constraint variable name together with the constraint variable name;
The relational expression is extracted from the constraint definition, and the extracted relational expression is converted into the internal expression specific to the constraint satisfaction solver by using the constraint variable converted into the internal expression specific to the constraint satisfaction solver stored in the constraint variable management unit. An application program execution device comprising: a constraint condition constructing means constructed as follows. A constraint satisfaction solution that satisfies a constraint definition consisting of a set of constraint conditions represented by a relational expression between constraint variables is specified from the data domain that is a set of possible values of the constraint variable. In a program for causing a computer to execute an application program obtained by using a constraint satisfaction solver that converts to an internal representation of
On the computer,
Reading the constraint definition from the constraint definition storage unit that stores the constraint definition handled by the application program ;
Extracting a constraint variable matching a condition of a predetermined constraint variable name from the constraint definition;
Acquiring the data domain of the constraint variable using the constraint variable name of the extracted constraint variable as a key from the data domain storage unit that stores the data domain handled by the application program together with the constraint variable name that is the name of the constraint variable;
Converting the obtained constraint variable in the data domain into an internal representation specific to the constraint satisfaction solver, and storing this together with the constraint variable name;
Extracting a relational expression from the constraint definition;
And executing the step of constructing the extracted relational expression as an internal representation specific to the constraint satisfaction solver using a constraint variable converted into the internal representation specific to the stored constraint satisfaction solver. Execution program.

Images