Jump to content
Sign in to follow this  
petrushka.googol

Theory of manpower sourcing for research projects

Recommended Posts

Whether this post belongs to the domain of motivation theory or compound probability or sociology or game theory or even abstract philosophy is hard to determine.

 

The gist of my argument goes thus "In a research project of n people with n1 people having a skill level of s1, n2 with s2...and so on", the optimum combination is a minimum number of n2 (where n2 > n1) with an exposure to the project lower than that of n1 (where exposure is measured in man-hours)", and others allocated along the same lines by extension of the same principle.

 

To illustrate consider a research project done by 3 individuals. Two of these have knowledge factors of 0.5 (an index of comfort level with the work done based on skill sets / experience etc.), and one has a knowledge factor of 0.8.

 

The collective efficiency of all 3 based on the above is 0.5 * 0.5 * 0.8 = 0.2

 

If all 3 were of the lower level of expertise their collective efficiency would be 0.5 * 0.5 * 0.5 = 0.125.

 

If all 3 were of the higher level of expertise their collective efficiency would be 0.8* 0.8 * 0.8 = 0.512.

 

No doubt the last option affords the best collective productivity (on an immediate basis) but sustaining this level over an extended period of time is more difficult. The second option on.the other hand offers less prospect of overall success.

 

The first option seems to afford the best mix of talent and expertise that is sustainable.

 

Thus selecting an ideal combination for a research endeavor is less likely to be a brute force combination of mediocre individuals or even highly talented individuals (who may be suspect to ego conflicts) and more a choice based on prudence as illustrated above.

 

Please express your viewpoints.

 

 

Share this post


Link to post
Share on other sites

Skil isn't a factor you can quantify like that.

If you have two men who can nearly, but not quite tie their shoelaces, there's no reason to suppose that they can collectively tie their laces.

Share this post


Link to post
Share on other sites

Skil isn't a factor you can quantify like that.

If you have two men who can nearly, but not quite tie their shoelaces, there's no reason to suppose that they can collectively tie their laces.

 

Skill level of a neophyte and an expert are not equal. That is an axiom.

I don't get the drift of the second argument.....That is exactly what i'm trying to say...mediocrity defeats purpose...introduce an expert into the group as an observer...albeit only for a short time...and he can transfer the knowledge required to tie the laces to the unskilled pair...which obviates the need for prolonged interaction with the control group once the primary objective is realized..and is in fact counter productive...for he / she could be used to transfer skills to another "compromised" pair....which is a win-win situation for all.

Share this post


Link to post
Share on other sites

There is no numerical measure of skill that you can do arithmetic with.

Share this post


Link to post
Share on other sites

Arbitrary arithmetics on top of that. Why are the values multiplied? Why are they not additive? Also according to that, it is always better to have individuals work on a given project as the addition of another person will always lower the score (assuming that the index is a value between 0 and 1).

Share this post


Link to post
Share on other sites

Arbitrary arithmetics on top of that. Why are the values multiplied? Why are they not additive? Also according to that, it is always better to have individuals work on a given project as the addition of another person will always lower the score (assuming that the index is a value between 0 and 1).

 

Adding probabilities gives total > 100% which is a fallacy. My tacit assumption is that if a task is assigned to 3 individuals it will be at least 3 times more complex and hence the minimum level of success is the product of their weighted averages. If the task is simple enough for 1 individual then the scenario is different and there is no need to add another person as you rightly mention.

Share this post


Link to post
Share on other sites

"Introduce an expert into the group as an observer...albeit only for a short time...and he can transfer the knowledge required to tie the laces to the unskilled pair."

So, he is a teacher.

But teaching is a different task and needs different skills.

I think many of us will have met people who can do their job, but are bad at explaining it to others.

Share this post


Link to post
Share on other sites

"Introduce an expert into the group as an observer...albeit only for a short time...and he can transfer the knowledge required to tie the laces to the unskilled pair."

So, he is a teacher.

But teaching is a different task and needs different skills.

I think many of us will have met people who can do their job, but are bad at explaining it to others.

 

Without resorting to apotheosis I would be tempted to use the euphemism "Consultant" ;)

Share this post


Link to post
Share on other sites

The most important in a research team (or indeed any team) is that the team has all the required skills available, and in the required amount (sometimes you need more than 1 person with the same skills), and that this team can cooperate as a team: The team will only be successful if the people can work together. In small teams, they can manage themselves, through efficient and timely communication. Larger teams require some form of leadership to make sure that people cooperate in a useful fashion.

 

Getting all the required technical skills into 1 team is easy. Getting people to cooperate is really really difficult (especially a bunch of really intelligent, but really stubborn researchers who are all too smart to think about mundane things like cooperation and teamwork :)). Unfortunately, in research, often the team-leaders (or managers) are people who used to be excellent at doing research, and were therefore promoted to lead the team. But leadership skills are something quite different than research skills.

Share this post


Link to post
Share on other sites

Adding probabilities gives total > 100% which is a fallacy.

You didn't give probabilities.

You gave skill level.

I guess so 1.0 means God-like know-everything.

 

Without actually knowing everything and having 1.0 skill, you can't even measure skills of current living people.

 

My tacit assumption is that if a task is assigned to 3 individuals it will be at least 3 times more complex and hence the minimum level of success is the product of their weighted averages.

 

You are oversimplifying.

They don't have to do the same task at the same time.

They can split job.

When one is f.e. figuring out spectral lines of Hydrogen, other one can work on Helium spectral lines, and yet another on Oxygen.

If all 3 would be trying to figure out spectral lines of one atom, their potential would be wasted, right. But on other hand, working on something doesn't necessarily means that there will be success.

 

Other person can inspire, encourage, motive to work even harder.

Give a new angle to look at problem when we stuck in it.

Show where has been made mistake or error.

 

If the task is simple enough for 1 individual then the scenario is different and there is no need to add another person as you rightly mention.

Your thread sounds to me a lot like discussion about making code multi-threaded.

There are tasks that are easy to split to multiple threads, multiple cpus,

and there are tasks that are hard,

and there are tasks that are impossible (the most likely they're serial dependent on previous calcs *).

 

Task easily splittable (scalable), can benefit nearly linearly from adding new resources.

Lost is only in merging/exchanging data.

 

*)

( a * b ) + c - impossible to split, adding c requires waiting for completion of a * b

( a * b ) + ( c * d ) - possible to split, a*b goes to 1st unit, c*d goes to 2nd unit, then final operation by one of them.

Edited by Sensei

Share this post


Link to post
Share on other sites

You didn't give probabilities.
You gave skill level.
I guess so 1.0 means God-like know-everything.

Without actually knowing everything and having 1.0 skill, you can't even measure skills of current living people.

 

You are oversimplifying.
They don't have to do the same task at the same time.
They can split job.
When one is f.e. figuring out spectral lines of Hydrogen, other one can work on Helium spectral lines, and yet another on Oxygen.
If all 3 would be trying to figure out spectral lines of one atom, their potential would be wasted, right. But on other hand, working on something doesn't necessarily means that there will be success.

Other person can inspire, encourage, motive to work even harder.
Give a new angle to look at problem when we stuck in it.
Show where has been made mistake or error.


Your thread sounds to me a lot like discussion about making code multi-threaded.
There are tasks that are easy to split to multiple threads, multiple cpus,
and there are tasks that are hard,
and there are tasks that are impossible (the most likely they're serial dependent on previous calcs *).

 

Task easily splittable (scalable), can benefit nearly linearly from adding new resources.

Lost is only in merging/exchanging data.

 

*)

( a * b ) + c - impossible to split, adding c requires waiting for completion of a * b

( a * b ) + ( c * d ) - possible to split, a*b goes to 1st unit, c*d goes to 2nd unit, then final operation by one of them.

 

To clarify :

 

Skill - level indicates possibility of succeeding in an atomic task with a level of understanding commensurate with the given weighted average.

 

I agree delegation is a factor but again, this only applies to individuals with equivalent levels of skill. You cannot delegate a task that requires basic skills to an expert because that would imply under-utilization of skills.

 

On the 3rd observation I concur, but then motivation is intangible and without that even forming a coherent team is impossible.

 

Multi-processing or multi-threading as you state implies asynchronous use of resources in a pool which is exactly what a team is like, if viewed as such. (mulit-processing does not imply multi-threading always). Serial tasks as you state cannot be delegated if they are dependent, but again what is the yardstick for atomicity? If you break up the first task into mutually exclusive sub-tasks assignment across resources becomes a realistic and feasible option.

 

Your thoughts.... :wacko:

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.