Better way to construct "sc concurrency"?

[xhoward]

Hi,

Attached model aims to model the planning phase of product development. I divide the initial "Tasks to Plan" into "int." and "sc." In the model I use a variable "sc concurrency" powered by a lookup with "%Tasks Released" by "Tasks to Plan(int.)" to control the availability of "Tasks to Plan(SC)."

Would you please help evaluate the adequacy of such an approach? Or is there any better alternative to measure concurrency and possibility to make the "sc concurrency" with randomness?

P.S.: I built for the model for 2 Taiwan high-tech companies in 2006. Though they are okay with the way I deploy the model, I still deliberate on more reasonable content.

Howard

[Edited on 4-9-2008 by xhoward]

[LAUJJL]

Hi Xhoward

Your question is not very clear.
It would be better to simplify your model to exclude everything not relevant to the problematic studied.
A simplistic model is easier to understand and avoids distraction. Simplifying the model to the utmost possible will help you too and will eventually increase the answers from the forum.
Regards.
JJ

[xhoward]

Hi JJ,

Thanks for your advice. I just posted a simplified model to clarify my question. In my model, the level of "Concurrence" (high or low) determines the availability of "Tasks to Do (B)," which subsequently affects the "Working Rate (B)".

I wonder if such definition of "Concurrence" is adequate and I would appreciate available comments. Please let me know if further explanations are required.

Howard

[LAUJJL]

Hi Xhoward

I have tried to reformulate your problem with plain english.

You have two stocks of tasks to accomplish in parallel A and B.

You try to accomplish FIRST the task A, taking into consideration
the labor productivity of A and the time to accomplish the task A.
The process limit A being the maximum tasks A that can be accomplished in a day
taking into account the duration of the work and the stock of A still to do.

You consider that at the beginning, the work generated by A leaves little time
to accomplish the tasks B by the effect of what you call concurrency.
In fact you do not consider the amount of work done for A as restricting the
possibility to work for B, but the amount of work still to be done for A.

Practically you consider that if there is still some work to do with A,
even if it is possible to do a certain amount of tasks in a day for B due to
the task process duration of B, you will not do it.

In fact in your model, everything is driven by the labour productivities
And it is therefore not possible to see the effect of the limiting action that
generates the non completion of task A on the rate of accomplishing task B.

I think that to help you a bit more, it is necessary to understand how the concurrency
effectively works and if your representation is sufficient to represent
the reality at least roughly. Does this model stands extreme situation or at
least reasonable conditions.

It would be good that you reformulate back the problem in plain English, including
all the implicit assumptions that you make and are not visible in your model.
Regards.
Joined a model with a slight modification to correct the unit error.
JJ

[xhoward]

Hi JJ,

I really appreciate your insights. I'll try my best to express my thoughts with plain English.

My modeling work has two primary goals. First, my client A wants to know how labor productivity affects the working rate. That is already exhibited in the model.

Second, because of the interdependencies of engineering tasks, task B cannot start to work until receiving information released from task A. This is to say that the completion of task A determines the availablity of tasks to be done by B. Since task B denotes tasks to be done by suppliers, client A wants to know if high concurrency, letting suppliers receive more information as early as possible, accelerates the whole process and eventually shortens project duration.

To make the concept of concurrency clearer, let me illustrate two examples. First, the completion of task A reaches 70%, and 30% of task B can be done. Second, the completion of task A reaches 30%, and 70% of task B is ready to begin. Obviously the concurrency of the previous is low and that of the latter is high.

My question is how to see the effect of the limited access to information released by A on accomplishing task B. Besides, I am not sure whether such definitions of concurrency are appropriate. Is there any better alternative to interpret and construct "concurrency"?

I uploaded a model with little revisions. Please let me know if any ambuiguity still exists.

Howard

[Edited on 4-11-2008 by xhoward]

[LAUJJL]

Hi Xhoward

The way you model the concurrency of two tasks is explained in Sterman Business dynamic at the end of chapter on lookup functions.
I have understood the principle and have to study again
the chapter I studied 6 years ago.
I have never modelled practical project models and I have no experience in that matter.
I think that if Sterman mentions that method, it is used.
I will have to think a bit to eventually find another way to represent the process.
The problem in that model is how you define the lookup
function. You can make a survey of how tasks A and B are related and build the lookup function from it.
Or you can get deeper in how the task A and B are related and describe it with a more sophisticated way than a simple lookup. The question too is why the user of the model is not totally satisfied by your model. By understanding the reason of it you may find some way to change the model
accordingly.
The administrator should know much better than me if there exists a better way to represent your problem
Regards.
JJ