JP2008152350A - Working time estimation device, working time estimation method, program and machine-readable recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the estimating precision of the time series distribution of working time in making a working plan in a software development project. <P>SOLUTION: A regression analysis part 2 performs the regression analysis of the working time data of a completed software development project with a time series model having model coefficients such as a Weibul model, and stores the model coefficients and a reuse ratio in a coefficient database 4 for every completion project. A coefficient estimation part 6 estimates the model coefficients of a new project from the reuse rate of the new project and the reuse rate of the completion project and the model coefficients. A total working time calculation part 7 calculates the working period and the total working time from the working start date and working completion date and scale of the new project. A time series distribution calculation part 8 calculates the time series distribution of the working time by using the time series model from the estimated model coefficients and the calculated working period and the total working time. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an operation time estimation device, an operation time estimation method, a program, and a machine-readable recording medium, and more particularly to an operation time estimation device, an operation time estimation method, a program, and a machine for managing the operation time of a software development project. It belongs to a readable recording medium.

  Conventionally, in order to improve the accuracy of work plan formulation of projects such as business and work for software development, various estimation methods used when formulating the work plan of the project have been proposed.

For example, Non-Patent Document 1 uses a model called a software development equation to estimate the total work time from the number of source code lines and the work period. A method for estimating the time series distribution of working hours using a time series model called the Norden-Rayleigh model has been proposed.
LHPutnam: “A General Empirical Solution to the Macro Software Sizing and Estimating Problem”, IEEE Trans. On Software Engineering, VOL.SE-4, No.4, pp.345-361, July 1978.

  In recent software development projects, new software is often created by diverting existing software in order to shorten the total work time or work period. In other words, the new software includes a part that is newly added to the diversion source software or a change in the diversion source software (referred to as a creation part) and a reuse part. The reusable part refers to a part that uses a part of the existing software of the diversion source as it is. As a result, in the software development, the work time required for design or implementation is reduced for the reusable part, and the ratio of the test work time to the total work time is increased. Since testing is mainly performed in the second half of the project period, the percentage of work hours in the second half of the project increases. Therefore, the time series distribution of the work time required for the software development project changes depending on the reuse rate indicating the ratio of the reused portion in the entire software to be developed.

  However, in the prior art as described in Non-Patent Document 1, the reuse rate is not considered when estimating the time series distribution of work time. For this reason, for example, when the conventional estimation method is applied to a project with a high reuse rate, the percentage of work hours in the first half of the project is overestimated, and the percentage of work hours in the second half of the project is underestimated. As a whole, the accuracy of estimation decreases. If a large error occurs in the estimated work time, a large error is included in the planned values such as the budget for the project and the number of workers, and even though the existing software was used to shorten the work period. Therefore, a great loss is incurred.

  SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a work time estimation device, a work time estimation method, a program, and a machine-readable recording medium on which the program is recorded, capable of estimating the time series distribution of the work time based on the reuse rate. Is to provide.

  According to one aspect of the present invention, a work time estimation device for estimating a work time related to a software development project using a computer having a processing unit and an output unit is a completed project in which the development of the software has been completed. Coefficient input means for inputting the reuse rate and model coefficient corresponding to the above, and main processing means for processing information on the new project scheduled to be developed for the new software.

  The model coefficient refers to a coefficient of the function, which is calculated by performing regression analysis on the work time data of each period of the completed project with a function having each period as a variable and having one or more coefficients, The reuse rate is based on the creation scale data indicating the scale of the created software among the completed software developed by the completed project and the reuse scale data indicating the scale of reuse of another existing software. Refers to the ratio of the reused size to the overall size of the completed software.

  Also, the reuse rate scheduled for the new project matches or approximates the reuse rate of the completed project.

  The processing means includes an estimated condition input unit for inputting a scheduled work start date, scheduled work completion date of the new project, scheduled creation scale data, and scheduled reuse scale data, and the estimated condition input. Based on the scheduled work start date, the scheduled work completion date, the creation scale data, and the reuse scale data input from the department, using the processing section, the total work according to the work period of the new project and a predetermined calculation formula Based on the total work time calculated by the total work time calculation unit, the work period, and the model coefficient, the processing unit is used to calculate the time of the new project. A time-series distribution calculating unit for calculating the work time of each period; and for each period, the work possible time scheduled for the new project and the previous time In association with the working time of the completion project calculated in time series distribution calculating section, and an estimation result output section for presenting via the output unit.

  Preferably, preprocessing means for processing information related to the completed project is further provided. The pre-processing means includes a project database storing work time data for each period of the completed project, the creation scale data and the reuse scale data for completed software developed by the completed project, By using the processing unit, the work time data of each period of the completed project read from the project database is subjected to regression analysis using a function having each period as a variable and having one or more coefficients. A regression analysis unit that calculates the model coefficient, which is a coefficient, and the entire completed software based on the creation scale data and the reuse scale data of the completed project read from the project database using the processing unit Said reuse rate, which indicates the ratio of reused size to size of Has a reuse rate calculating section for calculating the rate, the.

  Preferably, the regression analysis is performed in the case where work time data for each period, the creation scale data, and the reuse scale data are stored in the project database corresponding to each of a plurality of types of the completed projects. The unit uses the processing unit to perform regression analysis on the working time data read from the project database for each of the completed projects using a function having each period as a variable and having one or more coefficients. Thus, a model coefficient which is a coefficient of the function is calculated.

  The reuse rate calculation unit uses the processing unit to recycle the completed software with respect to the entire scale of the completed software based on the creation scale data and the reuse scale data read from the project database for each of the completed projects. Calculate the reuse rate that indicates the percentage of the scale used.

  The processing means calculates scheduled reuse of the new project based on the planned creation scale data and the planned reuse scale data input from the estimation condition input unit. Of the reuse rates calculated for the plurality of types of completed projects, the reuse rate that is greater than the scheduled reuse rate and most closely approximates the scheduled reuse rate, and the schedule The reuse rate that is smaller than the reuse rate that is approximated and that most closely approximates the planned reuse rate, and is selected according to the linear relationship indicated by the selected reuse rate and the corresponding model coefficient. A scheduled model coefficient calculation unit that calculates the model coefficient corresponding to the reuse rate.

  The time series distribution calculation unit uses the processing unit based on the total work time calculated by the total work time calculation unit, the work period, and the model coefficient calculated by the scheduled model coefficient calculation unit. Then, the working time for each period of the new project is calculated.

  Preferably, for each period, the workable time scheduled for the new project is calculated based on the number of scheduled workers in the period and the standard work time per worker.

  According to another aspect of the present invention, a work time estimation method for estimating a work time related to a software development project using a computer having a processing unit and an output unit is completed when the development of the software is completed A coefficient input step for inputting a reuse rate and a model coefficient corresponding to a project, and a main processing step for processing information on the new project scheduled to be developed for the software.

  The model coefficient refers to a coefficient of the function, which is calculated by performing regression analysis on the work time data of each period of the completed project with a function having each period as a variable and having one or more coefficients, The reuse rate is based on the creation scale data indicating the scale of the created software among the completed software developed by the completed project and the reuse scale data indicating the scale of reuse of another existing software. Refers to the ratio of the reused size to the overall size of the completed software, and the reuse rate scheduled for the new project matches or approximates the reuse rate of the completed project.

  The processing step includes an estimated condition input step for inputting a scheduled work start date, a scheduled work completion date of the new project, the planned creation scale data, and the planned reuse scale data, and the estimated condition input. Based on the scheduled work start date, the scheduled work completion date, the creation scale data, and the reuse scale data input in steps, the processing unit is used to perform the total work according to the work period of the new project and a predetermined calculation formula. Based on the total work time calculating step, the total work time calculated in the total work time calculation step, the work period, and the model coefficient, the processing unit is used to calculate the new project. A time-series distribution calculating step for calculating a working time for each period; and for each period, the new project in time series. Associating workable time to be scheduled and a working time of the completion project calculated in the time series distribution calculation step for, and an estimation result output step of presenting through the output unit.

  According to still another aspect of the present invention, there is provided a program for causing a computer to execute the above work time estimation method.

  According to still another aspect of the present invention, there is provided a machine-readable recording medium on which a program for causing a computer to execute the above work time estimation method is recorded.

  According to the invention, it is possible to estimate the time series distribution of the work time based on the reuse rate, so that it is possible to estimate with high accuracy.

Embodiments of the present invention will be described with reference to the drawings.
(Embodiment 1)
The work time estimation apparatus according to the present embodiment can be used for a software development project management system and the like.

  FIG. 1 shows a functional configuration of the work time estimation apparatus according to the present embodiment, and FIG. 2 is a block diagram of a computer on which the work time estimation apparatus according to the present embodiment is mounted.

  Referring to FIG. 2, the computer includes an input unit 700 having a keyboard 650 and a mouse 660, a central processing unit (CPU) 622 for centrally controlling and managing each part of the computer, a read only memory (ROM) or a RAM. (Random Access Memory) including a memory 624, a fixed disk 626, and an FD (Flexible Disk) 632 are detachably mounted, and an FD driving device 630 that accesses the mounted FD 632, a CD-ROM (Compact Disc) A read-only memory (642) is detachably mounted, and includes a CD-ROM drive device 640 for accessing the mounted CD-ROM 642, a communication interface 680 for communication connection between the communication line 300 and the computer, and a printer 690. . These units are connected for communication via a bus. Various wired or wireless lines such as the Internet can be applied to the communication line 300.

  Referring to FIG. 1, a work time estimation apparatus 100 includes a project database 1 and a coefficient database 4 corresponding to a memory 624 or a fixed disk 626, a regression analysis unit 2 realized as a function of the CPU 622, a reuse rate calculation unit 3, and a coefficient. Estimating unit 6, total work time calculating unit 7, time series distribution calculating unit 8, workable time calculating unit 10, estimated condition input unit 5 and resource condition input unit 9 realized using input unit 700, printer 609 or display An estimation result output unit 11 realized by using 610 is provided.

  The project database 1 stores scale data and work time data of completed projects. FIG. 3 is an example of the scale data 21 stored in the project database 1. FIG. 4 is an example of work time data 22 stored in the project database 1.

  Referring to FIG. 3, the project database 1 stores scale data 21 corresponding to each of a plurality of types of projects for which a software development project as a project has been completed. The scale data 21 includes project name data 211, which is information for identifying the project, and data indicating the scale of software that has been newly added or changed among the software completed by development (hereinafter referred to as update scale data). ) 212 and scale data 213 indicating the scale of the reused software. Note that KSLOC (1000 Software Lines Of Code) which is a unit of the scale data 212 and 213 indicates a scale indicating the number of source code lines in units of 1000 lines.

  Referring to FIG. 4, in project database 1, work time data 22 is stored corresponding to each of a plurality of types of software development projects that have already been completed. The work time data 22 includes project name data 221 which is information for identifying a corresponding project, a work date for software development of the project, a worker name and a work time of the worker corresponding to each work day. Work date data 222, worker name data 223, and time data 224.

  The regression analysis unit 2 obtains a model coefficient by performing regression analysis on the work time data 22 stored in the project database 1 using a time series model.

  The reuse rate calculation unit 3 calculates the reuse rate from the scale data 21 stored in the project database 1.

  The coefficient database 4 stores the model coefficient obtained by the regression analysis unit 2 and the reuse rate calculated by the reuse rate calculation unit 3. FIG. 7 is an example of coefficient data stored in the coefficient database 4.

  In the estimation condition input unit 5, the input unit 700 is operated by the person in charge of estimation, and data necessary for estimating a new project is input. FIG. 8 shows an example of input data input by the person in charge of quotation.

  The coefficient estimation unit 6 estimates (calculates) a model coefficient of a new project from the coefficient data stored in the coefficient database 4 and the input data input from the estimation condition input unit 5.

  The total work time calculation unit 7 calculates a work period and a total work time from the input data input by the estimation condition input unit 5 using existing technology.

  The time series distribution calculation unit 8 applies the model coefficient estimated by the coefficient estimation unit 6 and the work period and total work time calculated by the total work time calculation unit 7 to the time series model used by the regression analysis unit 2. Thus, the working time for each period is calculated. FIG. 10 is an example of work time data generated by the time series distribution calculation unit 8.

  In the resource condition input unit 9, the input unit 700 is operated by the person in charge of quotation and the resource condition of the new project is input. FIG. 11 shows an example of input data input by the person in charge of quotation.

  The workable time calculation unit 10 calculates the workable time for each period from the data input by the resource condition input unit 9. FIG. 12 is an example of workable time data generated by the workable time calculation unit 10.

  In the estimation result output unit 11, the data generated by the time series distribution calculation unit 8 and the workable time calculation unit 10 are output together to the display 610 or the printer 690 in a format such as a graph format. FIG. 13 shows an example of graph format output data output by the estimation result output unit 11.

  The processing of the work time estimation device 100 includes pre-processing for generating data to be stored in the coefficient database 4 based on data stored in the project database 1, data stored in the coefficient database 4, and an estimation condition input unit. 5 is divided into the main process for generating data to be output from the estimation result output unit 8 from the data input via 5.

  The pre-processing is started in response to an instruction given via the input unit 700 by the person in charge of estimation every time the completed project data set stored in the project database 1 is updated. This process is performed every time data is input from the estimation condition input unit 5 by the person in charge of estimation.

(Pre-processing)
First, the preprocessing operation will be described in detail with reference to FIG. 6 which is a flowchart showing the preprocessing operation. A program indicating the flowchart of FIG. 6 is stored in advance in the memory 624 and the like, and the processing of FIG. 6 is realized by the CPU 622 reading and executing each instruction code of the program. Here, it is assumed that the scale data 21 and the work time data 22 shown in FIG. 3 and FIG.

  In the pre-processing, when the person in charge operates the input unit 700 to instruct the update (change) of the data in FIG. 3 or 4, the CPU 622 reads the scale data 21 or work time data from the project database 1 in the memory 624 or the fixed disk 626. 22 is read, the read content is updated (changed) based on the data input from the input unit 700, and the changed data is stored in the project database 1 of the memory 624 or the fixed disk 626. Each time the data in the project database 1 is updated or every time the data is updated, an instruction is given via the input unit 700, so that the CPU 622 stores all the data stored in the project database 1 in accordance with the instruction. Start pre-processing for the project data.

  First, in step 101, the regression analysis unit 2 reads work time data 22 of one completed project from the project database 1. For example, it is assumed that the work time data 22 of the project Ti has been read. Subsequently, in step 102, the regression analysis unit 2 determines the work date of the data 222 indicating the earliest date among the work date data 222 of the read work time data 22 as the work start date.

  In step 103, the regression analysis unit 2 counts and accumulates the read work time data 22 for each date, and generates, for example, cumulative work time data 23 as shown in FIG. The accumulated work time data 23 includes project name data 231, work day data 232, elapsed days data 233 and accumulated time data 234 corresponding to each work day. The elapsed days data 233 indicates the elapsed days from the work start date determined as described above to the corresponding work day. The accumulated time data 234 indicates the accumulated value of the work time from the work start date determined as described above to the corresponding work date. The accumulated work time data 23 generated for each project is stored in the memory 624 or the like.

  In step 104, the regression analysis unit 2 processes the accumulated work time data 23 read from the memory 624, and among the dates indicated by the work day data 232, for example, the time indicated by the corresponding accumulated time data 234 is 99. The date of 9% is detected and determined as the work completion date. Specifically, the cumulative time data 234 corresponding to the data 232 indicating the final work day among the work day data 232 of the cumulative work time data 23 is read as the total work time, and is 99.9% of the read total work time. The corresponding time is calculated, the calculated time is compared with each of the accumulated time data 234, data 234 indicating the accumulated time closest to the calculated time is detected based on the comparison result, and the detected accumulated time data 234 is detected. The work day data 232 corresponding to is read from the accumulated work time data 23. The date indicated by the work date data 232 read in this way is determined as the work completion date of the project.

  Next, in step 105, the regression analysis unit 2 calculates the work period by calculating the difference between the work start date determined in step 102 and the work completion date determined in step 104.

  In step 106, the regression analysis unit 2 performs a regression analysis of the accumulated work time data 23 using a time series distribution model having at least one model coefficient such as a Weibull model, and calculates the model coefficient. The Weibull model is a time series model expressed by Equation 1.

  Equation 1 above has two model coefficients, α and β. Here, t corresponds to an explanatory variable in the regression analysis and indicates an elapsed period. The variable K (t) corresponds to the objective variable (dependent variable) in the regression analysis and indicates the accumulated work time in the elapsed period t. The variable K indicates the total work time, and the variable Td indicates a constant work period. Therefore, the variable K and the coefficients α and β are values to be calculated by regression analysis.

  Here, by setting the date when the accumulated work time becomes 99.9% of the total work time in step 104 as the work completion date, K (t) /K=0.999 at t = Td. Therefore, Equation 1 is converted to Equation 2 below.

  In Equation 2, the coefficient β is shown as a function of the coefficient α. Therefore, when the value of the coefficient α is determined, the value of the coefficient β is automatically determined (calculated). The data of the calculated coefficient α is stored in the memory 624.

  In step 107, the reuse rate calculation unit 3 reads the scale data 21 of the same project as the work time data 22 read in step 101 from the project database 1. Here, since the working time data 22 of the project name Ti is read in step 101, the scale data 21 in which the project name data 211 indicates “Ti” is read from the project database 1.

  In step 108, the reuse rate calculation unit 3 calculates a reuse rate indicating the ratio of the number of source code lines reused to the number of source code lines in the developed software according to Equation 3. The calculated reuse rate data is stored in the memory 624 or the like.

  In step 109, the regression analysis unit 2 and the reuse rate calculation unit 3 read the model coefficient α and the reuse rate data calculated in step 106 and step 108 from the memory 624, and associate them with the project name data in the model coefficient. And the reuse rate data are stored in the coefficient database 4.

  Referring to FIG. 7, in the coefficient database 4, the project name data 241 corresponding to each project, the model coefficient α calculated by the regression analysis unit 2 and the reuse rate calculation unit 3 and the reuse rate for the project are stored. Data 243 and 242 are stored.

  In step 110, the regression analysis unit 2 determines whether the pre-processing in steps 101 to 109 has been performed for all projects whose data is stored in the project database 1. If it is determined that all have been performed (“Yes” in Step 110), the pre-processing is terminated, and if not (“No” in Step 110), Step 101 is performed on the remaining project data in the project database 1. The process in step 109 is repeated.

  The determination of the end in step 110 is performed by, for example, counting the total number of scale data 21 or work time data 22 stored in the project database 1 in advance, and executing the processing in steps 101 to 109 for the number of count values. When it is done, it may be determined to end. Note that the determination method is not limited to this.

  When the scale data 21 and the work time data 22 shown in FIGS. 3 and 4 are stored in the project database 1, after the pre-processing shown in FIG. 6, the coefficient data shown in FIG. Is memorized.

(This processing)
Next, the operation of this process will be described in detail with reference to FIG. 9 which is a flowchart showing the operation of this process. 9 is stored in advance in the memory 624 and the CPU 622 reads out and executes each instruction code of the program, thereby realizing the processing in FIG. Here, at the start of this processing, the scale data 21 and the work time data 22 shown in FIGS. 3 and 4 are stored in the project database 1 in advance, and the data shown in FIG. 7 is stored in the coefficient database 4. Assume that

  In step 201, the person in charge of the estimate operates the input unit 700 and inputs, for example, condition data 25 as shown in FIG. 8 via the estimate condition input unit 5. The condition data 25 is stored in the memory 624.

  The condition data 25 is data related to a newly planned software development project, and indicates the condition for estimating the work time. Specifically, the start date and the completion date of work of the project are indicated. Data 251 and 252 to be pointed out, data indicating the scale of software to be newly added or changed among the software to be completed (hereinafter referred to as update planned scale data) 253, and a reuse schedule indicating the scale of software to be reused Scale data 254 is included. The unit of the scale data 253 and 254 is KSLOC.

  In step 202, the coefficient estimation unit 6 reads the condition data 25 from the memory 624, and uses the update scheduled scale data 253 and the reuse scheduled scale data 254 of the read condition data 25 according to the same calculation method as in the preprocessing step 108. Calculate the expected reuse rate (%). For example, in the case of the condition data 25 in FIG. 8, since the update planned scale data 253 and the reuse planned scale data 254 are 1000 (KSLOC) and 4000 (KSLOC), respectively, the planned reuse rate is 4000 / (1000 + 4000). X100 = 80.0 (%) The calculated reuse rate is stored in the memory 624 as scheduled reuse rate data 26.

  In step 203, the coefficient estimation unit 6 calculates (estimates) a model coefficient for a project related to new software. First, the scheduled reuse rate data 26 calculated in step 202 is read from the memory 624, and the coefficient database 4 is searched based on the reuse rate indicated by the read scheduled reuse rate data 26. Specifically, the data 242 closest to the scheduled reuse rate data 26 among the data 242 indicating the reuse rate having a value larger than the reuse rate indicated by the scheduled reuse rate data 26, and the value smaller than that. Of the data 242 indicating the reuse rate, the data of two projects, the data 242 closest to the planned reuse rate data 26, are read from the coefficient database 4. For example, in the coefficient database 4 of FIG. 7, the data of the project Pa with a reuse rate of 95.2 (%) and the data of the project Ti with a reuse rate of 72.0 (%) are read out. Here, it is assumed that the data 242 indicating the same value as the calculated scheduled reuse rate data 26 is not stored in the coefficient database 4.

  In step 204, the coefficient estimation unit 6 estimates the model coefficient α at the reuse rate of 80.0 (%) by linearly interpolating the data of the two projects Pa and Ti read in step 203. Specifically, since the model coefficient when the reuse rate is 95.2% is 6.42 and the model coefficient when the reuse rate is 72% is 5.24, the reuse rate is x (% ) And the model coefficient can be expressed as a linear relationship of y = 0.0509x + 1.5779. Substituting the target reuse rate of 80 (%) into this relational expression yields y = 5.6499≈5.65. By performing linear interpolation in this way, the model coefficient α for a project having a reuse rate of 80.0 (%) can be estimated (calculated) as about 5.65.

  Note that the reuse rate and model coefficient do not have a linear relationship over the entire range, but are considered to have the above-described linear relationship between two points in the vicinity of the target reuse rate. There is no problem. Therefore, in the present embodiment, the model coefficient is estimated (calculated) according to the above-described linear relational expression defined by these two points.

  In step 205, the total time calculation unit 7 reads the data 251 and 252 from the condition data 25 (stored in the memory 624) input in step 201, and the scheduled work start date and scheduled work completion date indicated by these read data. The scheduled work period 28 is calculated by calculating the difference between and stored in the memory 624. For example, in the case of the condition data 25 in FIG. 8, the scheduled work period 28 because the scheduled work start date indicated by the data 251 and 252 is September 21, 2006 and the scheduled work completion date is December 25, 2006. Is calculated as 96 (days).

  In step 206, the total work time calculation unit 7 calculates the addition or change scheduled scale indicated by the data 253 and 254 of the condition data 25 input in step 201, the planned reuse period, and the scheduled work period 28 calculated in step 205. Based on the existing technology, the total work time is calculated using the following equation 4 indicating the software development equation shown in Non-Patent Document 1, for example. Here, the total work time K is calculated by applying Equation 4 to each of the additional and changed scales and the reused scale.

Here, the variable Ss indicates the total number of source code lines, the variable K indicates the total work time to be calculated, and the variable t _d indicates the work period 28 that is a constant. The variable Ck indicates a predetermined productivity coefficient, and a different coefficient value is applied to each of the scale that has been added and changed and the scale that has been reused.

  In step 207, the time series distribution calculation unit 8 calculates the model coefficient estimated in step 204 to the same time series distribution model used in step 106 of the preprocessing, the scheduled work period calculated in step 205, and the calculation in step 206. The accumulated work time with respect to the elapsed days is calculated by applying the total work time thus obtained. In addition, the work time of each work day is calculated (estimated) by calculating the difference in the accumulated work time of the work days. Thus, for example, work time data 30 as shown in FIG. 10 is generated and stored in the memory 624.

  The work time data 30 in FIG. 10 includes work day data 301, data 302 indicating the number of days elapsed from the work start date for each work day, and data 303 indicating the accumulation of work time from the work start date to the work day. And data 303 indicating the work time of the work day.

  In step 208, the person in charge of quotation operates the input unit 700 to input resource condition data 31 as shown in FIG.

  The resource condition data 301 includes work day data 311, data 312 indicating the number of workers scheduled for each work day, and data 313 indicating the standard work time per worker.

  In step 209, the workable time calculation unit 10 calculates the product of the planned number of workers on each work day input in step 208 and the standard work time, for example, workable time data 32 as shown in FIG. Generated and stored in the memory 624.

  The workable time data 32 in FIG. 12 includes workday data 321 and data 322 indicating workable time for each work day. The workable time data 322 is indicated by data 312 indicating the number of workers on the corresponding work day × data 313 indicating the standard work time per worker.

  In step 210, the estimation result output unit 11 associates the estimated work time data 30 generated in step 207 with the workable time data 32 generated in step 209, and displays the display 610 in a graph format as shown in FIG. Are output to the screen or the printer 690.

  In FIG. 13, the date indicated by the work date data 301 and 321 is taken on the horizontal axis of the graph, and the work time is taken on the vertical axis. Corresponding to each date on the horizontal axis, the work time data 303 pointed to by the estimated work time data 30 is a bar graph, and the work time data 322 pointed to by the work time data 32 is a line graph associated with the time series according to the work date. Is shown.

  The person in charge of estimation can confirm the difference between the work time and the workable time for each work day by checking the output data as shown in FIG. 13 output by the estimation result output unit 11. For example, in FIG. 13, the second half of the work period indicated by the horizontal axis corresponds to a software test period, and the first half corresponds to a period required for design or implementation. From the graph, it can be understood that the number of scheduled workers may be reduced because the work time that can be allocated in the period of design or implementation is longer than the estimated work time. Also, there is a part in the test period that exceeds the estimated work time that can be spent, so this part increases the planned number of workers, for example, the surplus in the first half period It can be easily understood that the work time should be moved to the second test period.

(Other examples)
In FIG. 1, the coefficient estimator 6 estimates (calculates) the model coefficient by linear interpolation. However, the coefficient may be set as follows, that is, the reuse rate scheduled for a new project is completed. If the project reuse rate matches or approximates, the model coefficient calculated by the regression analysis unit 2 for the completed project may be used without using linear interpolation. In this case, the model coefficient and the reuse rate for the completed project are input from the input unit 700.

(Embodiment 2)
In the present embodiment, a program for realizing the work time estimation apparatus 100 of the first embodiment is stored in a computer-readable recording medium in FIG.

  As a general tendency, various program modules are prepared as a part of a computer operating system, and an application program is called to advance the processing by calling these modules in a predetermined arrangement. . In such a case, the software itself for realizing the work time estimation apparatus 100 does not include such a module, and the work time estimation apparatus 100 is realized only when the computer cooperates with the operating system. However, as long as a general platform is used, it is not necessary to distribute software including such modules, and the software itself not including these modules and the recording medium storing the software (and the software distributes on the network). Data signal) can be considered to constitute the embodiment.

  In the present embodiment, as the recording medium, a memory required for processing by the computer shown in FIG. 2, for example, the memory 624 itself may be a program medium. Alternatively, program reading devices such as an FD driving device 630 and a CD-ROM driving device 640 are provided as external storage devices, and FD 632 and CD-ROM 642 as storage media are respectively inserted into the program reading devices. There may be. In any case, the stored program may be configured to be accessed and executed by the CPU 642, or in any case, the program is read once, and the read program is the predetermined program shown in FIG. A method may be used in which the program is loaded into the program storage area and read and executed by the CPU 642.

  Here, the program medium described above is a recording medium configured to be separable from the computer main body, and is a tape system such as a magnetic tape or a cassette tape, a magnetic disk such as an FD, or a CD-ROM 642 / MO (Magnetic Optical Disc) / MD. Optical discs such as (Mini Disc) / DVD (Digital Versatile Disc), card systems such as IC cards (including memory cards) / optical cards, mask ROM, EPROM (Erasable and Programmable ROM), EEPROM (Electrically EPROM) ), Or a medium carrying a fixed program including a semiconductor memory such as a flash ROM.

  2 employs a configuration that can be connected to the communication line 300 including the Internet via the communication interface 680. Therefore, a medium that fluidly carries the program by downloading the program from these communication networks. It may be.

  Note that the content stored in the recording medium is not limited to a program, and may be data.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a functional block diagram of the working time estimation apparatus which concerns on this Embodiment. It is a block diagram of the computer which concerns on this Embodiment. It is a figure which shows an example of the scale data in a project database. It is a figure which shows an example of the working time data in a project database. It is a figure which shows an example of accumulation work time data. It is a flowchart which shows the operation | movement of a pre-processing. It is a figure which shows the example of the content of a coefficient database. It is a figure which shows an example of condition data. It is a flowchart which shows operation | movement of this process. It is a figure which shows an example of work time data. It is a figure which shows an example of resource condition data. It is a figure which shows an example of work possible time data. It is a figure which shows the example of the content output by the estimation result output part.

Explanation of symbols

  1 Project database 2 Regression analysis unit 3 Reuse rate calculation unit 4 Coefficient database 5 Estimation condition input unit 6 Coefficient estimation unit 7 Total work time calculation unit 8 Time series distribution calculation unit 9 Resource condition input unit 10 Workable time calculation unit, 11 Estimation result output unit, 100 Worktime estimation device.

Claims (7)

A work time estimation device for estimating a work time related to a software development project using a computer having a processing unit and an output unit,
Coefficient input means for inputting a reuse rate and a model coefficient corresponding to a completed project for which development of the software has been completed,
And processing means for processing information related to the new project scheduled to be developed for the new software,
The model coefficient refers to a coefficient of the function, which is calculated by performing regression analysis on the work time data of each period of the completed project with a function having each period as a variable and having one or more coefficients,
The reuse rate is based on creation scale data indicating the scale of created software among completed software that has been developed by the completion project, and reuse scale data indicating the scale of reuse of another existing software, Refers to the ratio of the reused size to the overall size of the completed software;
The reuse rate scheduled for the new project matches or approximates the reuse rate of the completed project;
The main processing means includes
An estimated condition input unit for inputting the scheduled work start date, scheduled work completion date of the new project, the planned creation scale data, and the planned reuse scale data;
Based on the scheduled work start date, the scheduled work completion date, the creation scale data, and the reuse scale data input from the estimation condition input unit, the work unit and predetermined calculation of the new project are performed using the processing unit. A total work time calculation unit for calculating a total work time according to the formula;
Time series distribution calculation for calculating the work time of each period of the new project using the processing unit based on the total work time, the work period, and the model coefficient calculated by the total work time calculation unit And
Estimates to be presented via the output unit in association with the workable time scheduled for the new project and the work time of the completed project calculated by the time series distribution calculation unit for each period in time series A work time estimation device having a result output unit. Further comprising pre-processing means for processing information relating to the completed project;
The pre-processing means includes
Project database storing work time data for each period of the completed project, the creation scale data and the reuse scale data for the completed software developed by the completed project,
By using the processing unit, the work time data of each period of the completed project read from the project database is subjected to regression analysis using a function having each period as a variable and having one or more coefficients. A regression analysis unit for calculating the model coefficient, which is a coefficient;
Based on the creation scale data and the reuse scale data of the completed project read from the project database by using the processing unit, the reuse indicating the ratio of the reused scale to the entire scale of the completed software. The work time estimation apparatus according to claim 1, further comprising: a reuse rate calculation unit that calculates a utilization rate. In the project database, corresponding to each of a plurality of types of the completed projects, when working time data for each period, the creation scale data, and the reuse scale data are stored,
The regression analysis unit
By using the processing unit, for each of the completed projects, by performing regression analysis on the work time data read from the project database with a function having each period as a variable and having one or more coefficients, Calculating a model coefficient which is a coefficient of the function;
The reuse rate calculation unit
Using the processing unit, for each of the completed projects, based on the creation scale data and the reuse scale data read from the project database, the ratio of the reused scale to the overall scale of the completed software Calculate reuse rate,
The main processing means includes
A scheduled reuse rate calculation unit that calculates the scheduled reuse rate of the new project based on the scheduled creation scale data and the scheduled reuse scale data input from the estimation condition input unit;
Of the reuse rates calculated for the plurality of types of completed projects, the reuse rate that is greater than the scheduled reuse rate and is closest to the scheduled reuse rate, and the scheduled reuse rate The reuse rate that is smaller than and most similar to the scheduled reuse rate, and follows the linear relationship indicated by the selected reuse rate and the corresponding model coefficient to the scheduled reuse rate. A planned model coefficient calculation unit for calculating the corresponding model coefficient,
The time series distribution calculation unit uses the processing unit based on the total work time calculated by the total work time calculation unit, the work period, and the model coefficient calculated by the scheduled model coefficient calculation unit. The work time estimation device according to claim 2, wherein the work time for each period of the new project is calculated.   The workable time scheduled for the new project for each period is calculated based on the number of scheduled workers in the period and the standard work time per worker. Work time estimation device described in 1. A work time estimation method for estimating a work time related to a software development project using a computer having a processing unit and an output unit,
A coefficient input step for inputting a reuse rate and a model coefficient corresponding to a completed project for which development of the software has been completed;
And processing step for processing information on the new project scheduled for development of the software,
The model coefficient refers to a coefficient of the function, which is calculated by performing regression analysis on the work time data of each period of the completed project with a function having each period as a variable and having one or more coefficients,
The reuse rate is based on creation scale data indicating the scale of created software among completed software that has been developed by the completion project, and reuse scale data indicating the scale of reuse of another existing software, Refers to the ratio of the reused size to the overall size of the completed software;
The reuse rate scheduled for the new project matches or approximates the reuse rate of the completed project;
The main processing step includes
Estimated condition input step for inputting the scheduled work start date, scheduled work completion date of the new project, scheduled creation scale data, and scheduled reuse scale data;
Based on the scheduled work start date, the scheduled work completion date, the creation scale data, and the reuse scale data that are input in the estimation condition input step, the processing unit is used to calculate the work period of the new project and a predetermined calculation. A total work time calculating step for calculating a total work time according to the formula;
Time series distribution calculation for calculating the work time of each period of the new project using the processing unit based on the total work time, the work period and the model coefficient calculated in the total work time calculating step Steps,
Estimates to be presented via the output unit in association with the workable time scheduled for the new project and the work time of the completed project calculated in the time-series distribution calculation step for each period in time series A work time estimation method comprising a result output step.   A program for causing a computer to execute the work time estimation method according to claim 5.   A machine-readable recording medium recording a program for causing a computer to execute the work time estimation method according to claim 5.

Images