Welcome to free paper download website

Economic other

You are here: Home > Economics > Economic other > content

The software project schedule management on system dynamics

Author: HeManHui YangZuoPing From: www.yourpaper.net Posted: 2010-05-29 04:42:56 Read:

Abstract: uncertainty and nonlinear complex software project management, especially in the schedule management.Method: the article uses system dynamics simulation and implementation process of the two stages of software project, the model of the project completion time can effectively predict the influence of the staff, at the same time distribution on the progress of the project are also discussed.
Keywords: the software project management; project management; system dynamics; prediction; staff assignments

In order to solve the "software crisis", the academic circles and the theory of project management experience in software development, the birth of the software project management.As everyone knows the schedule management, quality management and cost management as the main content of software project management.In schedule management, generally adopt the traditional project management methods, such as the Gantt chart, critical path method and program evaluation and review technique, they have established in the project can be divided into separate processes, but in the actual software project management, each stage (processes) are interconnected, like the previous stage not found errors will implement a stage to later stage, at the same time when found a mistake to a stage of rework, and so on so many phenomenon, these interactions are often nonlinear, which in the traditional network diagram is difficult to express, is beyond the scope of the manager's mind to understand.
System dynamics, Massachusetts Institute of Technology professor Forrester nineteen fifties founded provides tools and a feasible theory, idea, method for solving the dynamic complex problems.
The 1 system dynamics
Nineteen fifties Forrester professor of computer science and the feedback control theory is applied to society, economy and system research.
In twentieth Century, 70 years, system dynamics has gradually become a kind of understanding and knowledge of research methods of complex system of human dynamics.2O years scholars like Professor Wang Qifan participated in the application of system dynamics research work in China, and has made important contributions.
System dynamics in software project management application is relatively rare, estimate time and cost are discussed by using the method of system dynamics, the system dynamics method to study the influence of schedule target on project performance.We discuss the use of system dynamics method in project risk management, especially in the management of project risk dynamics direction features and advantages.
System dynamics method emphasizes the view of closed loop to identify and solve problems, which also determines its modeling method using feedback loop, namely through the feedback loop structure of the back analysis of behavior mode, change the value in the relevant variable structure, to understand the different behaviors under different tactics, to complete the optimization strategy.
System dynamics that influence structural relationship, feedback loop time delay, information amplification behaviour of the system, the structure relationship indicates that relations between the structure of the system, the time delay said that the decision action behind the information, information amplification expressed with process and the passage of time, some information will be enlarged, its influence on decision-making behavior will be magnified.
2 model
Software project management often includes a plurality of stages, these stages are interrelated, mutual network.But the relationship between two stages is the foundation, the 2 phase of software project development as an example.The model based on the following assumptions: each stage of development and there are some errors, these errors are a part at this stage was correct, part of a need to the next stage in order to find.Discovery of error probability and engaged in the debugging personnel in each stage of.Error effect of upper stage left the next phase of development and debugging.In each stage of the staff is fixed, the development will inevitably lead to fewer number of debugging.Debugging the project schedule for the first stage of development, second stages of development, the first phase of the second stage, rework, debugging.The first stage and the second stage is the rework debugging parallel.
> connection
The model state variables of a total of 5, "hair development in 1", "1" project has changed ", reworked items", "hair development in 2", has "change the project 2" respectively, the first phase of the development of debugging, rework, second stages of development, debugging, the relationship in figure 1.


Only these 5 procedures are completed, the project is completed, 2 auxiliary variables "the whole project" said the project progress of concern.


2.2 rate variables to set the
The model includes 5 rate variables, which is arranged respectively: the actual development rate of 1= remaining work effect of 1 กม 1 กม development rate development efficiency * development number 1.
A number of 1 กม 1 debugging debugging debugging efficiency x remaining errors affect 1 กม time between 1.
The actual development rate 2 * 2 * development efficiency remaining work effect of 2 * time between 2.
A number of 2 กม 2 debugging debugging debugging efficiency x remaining error on phase error of 2 * * time between 3 rework = debugging number 1 กม debugging efficiency x remaining legacy of errors.
Need explanation is: "remaining work effect", "residual error effect", "the remaining legacy of errors".The three main uses in Figure 3 reference model.

"The remaining work (error) effect" that when work started working rate is relatively low, when the working end speed rate is relatively low, intermediate the fastest rate; "surplus left mistakes" said as the first stage of the legacy of the more mistakes, the second phase of the slower; this is consistent with the actual situation.
Other auxiliary variable settings will not go into the details, the model of the whole flow diagram is shown in figure 4.



Estimation of 3 project schedule
Set up the model, by setting the constants and variables can be used to simulate the system.Here the first and second stage of the staff of 20 people, the first stage and the second stage is expected workload is 1000, the error rate is 0.2, the developer was 10 at the first stage, commissioning personnel 10, second stage development staff to 7, commissioning personnel 13, simulated project schedule as shown in Figure 5 show.


From Figure 5 can be seen in the whole project 160 (days), and thirty-fifth (day) the first stage first developed, debugged in 70 days, to complete the second phase of the initial development in 125 days, the first phase completed debugging rework and second stages in 160 days.
4 personnel distribution / > effect of

From the model flow chart (Figure 4) can be seen, the decision variables are allocated to each stage of personnel and personnel's work efficiency.Improve work efficiency will accelerate progress, it does not need to discuss, here is the main research personnel allocation, especially the influence of personnel allocation the first phase of the entire project plan.
Model hypothesis first stage left to the second stage of the error depends on the first stage in the debugging stage number, namely: 1 left wrong wrong (wrong * debugging number 1 role).
The "debug number 1 role" adopted "S curve" reference model, said debugging are few in number and found the error probability is small, with the increase in the number of discovered quickly upgrade the error probability, the number of saturated probability tends to be stable.
In the first stage of the second phase of the remaining errors, to second
Effect of the debugging process stage as an example to illustrate the model assumptions and debugging, in the second stage can be gradually discovered by the error on the first stage, therefore, to the first stage of rework, and commissioning of second stage was continued, and the debugging speed is the first phase of the remaining legacy errors, namely: a number of 2 กม 2 debugging debugging debugging efficiency x remaining error on phase error of 2 * * time between 3.

Figure 6 was used to simulate the first stage developers premise of total staff to 20 people for 5, 10, 15, 18, a total of four cases corresponding to the progress of the whole project.From the chart we can see that the developer is 15 faster, developers too much (18) and low (5 people) will lead to an increase in the progress of the project, which is consistent with the reality.Result is that such principle: when the developer is too small, then the development process; when too many developers, will inevitably lead to the debugging personnel too little, although the development of the first phase of the process time is shortened, but is bound to increase the debugging process time, and increase the remaining errors and then affect the next stage of progress.Therefore the personnel allocation scheme for a moderate.
 1/2    1 2 Next Last
Please consciously abide by Internet-related policies and regulations.
Tips: Log in to comment, the user name to enter comments directly from your personal space, so that more friends to meet you.

Economic other latest papers

Sponsored Links

Economic other papers Ranking

Latest free papers

Sponsored Links

Top