Can Matlab help in design experiments?

I have four different type of inputs and three type of outputs, I want to design my experiments with these values in Matlab and I don't know how can I do it? My inputs are Continuous values as given below:

I1=[1.5,5.5]mm

I2=[5,85]degree

I3=[100,300]mm

I4=[5,50]mm

And Outputs ranges are unkown:

O1=[?,?]m/s

O2=[?,?]pa

O3=[?,?]kg

I need two methods to design my experiments with continuous values and discontinuous values whether I prefer to write Input ranges in discrete numbers like:

I1=[1.5, 2.5, ... 5.5]mm

I2=[5, 10, 15,... 85]degree

I3=[100, 150, 200, .... 300]mm

I4=[5, 10, 15, ... 50]mm

[Adam]

Working with continuous values is not possible if you want to use all values in the continuous range. If you wish to just pick one value from each continuous input for an experiment then that is, of course, possible.

It depends what your experiments involve though. Do you have a formula that gives O1, O2 and O3 based on I1, I2, I3 and I4?

[Ayob]

I don't have any formula.Actually these input valuse I1, I2, I3 and I4 are four different geometric parameters in my project (My project is model of a real machine). Since it's not possible to do modeling for all of them (I1*I2*I3*I4), I need a summary of them to use in my modelings and train genetic algorithm. There is no problem, I can use discontinuous values of these ranges( I1,I2 ...I4)

[dpb]

An experiment for what purpose? Optimal design depends on the null hypothesis one is trying to show or perhaps you're trying to do a response surface to build a model?

[Ayob]

Actually number of modelings should be summarized.Since for a complete analysis I need a large number of modelings (I1*I2*I3*I4=number of modelings) and it's not possible to do such a heavy work for me. so I need to summrize number of modelings and use just some modelings.

[dpb]

I have no clue what you're trying to say here...again, I repeat that a design is dependent upon the objective. Matlab Statistics Toolbox has some D(esign)OfE(xperiment) tools, yes, but what to use and how is dependent upon the objective of the experiment.

Amplify/explain what you think you're trying to get as an end result after you run the experiment (whatever the final design) and that'll at least give us some clues as to where to point.

ADDENDUM

Perhaps you simply mean you want a fewer number of design point; yes, that's part of the purpose of DOE in aiding the selection of experimental points to provide more information at less cost. The larger thing they can provide, however, is the assurance that in the end one can actually estimate the desired effects or coefficients whether the design is truly "optimal" or not.

[Ayob]

Yes, exactly. I need help to find the appropriate codess in Matlab.

(Are there codes for working with continuous values which are introduced in I1,I2..I4 ranges?)

[dpb]

Yes, but still it's the objective of the experiement that is the key, first before one can decide what is the appropriate design. Hence the tool selection follows that.

The tool can't make the decision as to what it is that is the purpose of running the experiment in the first place; only your (apparently a well-kept secret as you don't seem willing to share even a hint) knowledge of what you're trying to accomplish as the end result can define that.

First, you need to write the null hypothesis of the test and decide the proper analysis technique that will let you compute and estimate the necessary statistics to test that hypothesis or if it is a model you're trying to build have enough background information on the process to be able to select an appropriate one and then be able to estimate the underlying coefficients.

You're putting the cart before the horse by, it seems, expecting one of these tools to do that for you magically.