How to group the corresponding elements of an array?

Hi,

I'm calculating the value of a parameter f_x whose value changes according to the following conditions.

I haven't shown the equation for all the parameters here, as this would be enough to give an idea on how f_x is calculated and how it's value is changing according to the above conditions.

Now my question is how will I calculate f_x by grouping all the 5 parameters values separately for dry, wet, low and very low conditions.

For eg.,I need to give something like this,

 f_x.dry= F_x/Q; 
f_x.wet=F_x/Q;

where f_x gives me a value on considering the dry condition data. (Likewise for other conditions). So how would I need to define the conditions, so that I can give f_x.dry

Any help would be greatly appreciated.

Thanks

[Sara]

Are you familiar with .* and ./? they allow you to multiply and divide arrays element by element (in contrast with matrix algebra). If you use that notation in your equations, then f_x will be a four element vector, one element for each conditions (dry, wet, low, very low). Since you have not provided a full code, it's difficult to comment on the specifics.

[Priya]

Thanks. I need to check this out and if it's not going to work, then will it be fine to post my entire code?

[Sara]

Sure, posting a code is fine. Just post all of it, inputs included.

[Sara]

First, you'll need to tell us the value of these variables:

E_r poi_rat_r E_w poi_rat_w Q

Second, matlab is case sensitive and a is not the same as A. In the code above, you have a and b, did you mean A and B?

Third, you can't use [] brackets unless you are trying to concatenate something (e.g., two arrays). If that's the math notation, use () in matlab.

Forth, you have ^3/2 on a line. Did you mean to elevate at the 1.5 power of at the third power and then divide the results by two?

[Joseph Cheng]

I agree with Sara's 4th point. x^3/2 is not to the 3/2 power is actually (x^3)/2. However i learned something today. [] does work as it should in math notation. However i can see the downsides of interchanging its useage as it could be confused with array concatenation.

in summary are you attempting to run this 4 times for each of the conditions? where the end results will give you either f_x.dry or f_x.(condition) with the parameters corresponding to the condition variable? if so why not just use a for loop to loop when condition=1:4 and a switch case statement based on condition.

for condition = 1:4
     k_A = k_A_index(condition )
     k_S = k_S_index(condition )
     cof_0 = cof_0_index(condition )
     A = A_index(condition )
     B = B_index(condition )
 %%do processing stuff here

switch condition 
case 1
f_x.dry = ________________
case 2
fx_x......

end
end

[Priya]

Thanks very much. Could you please tell me how to define the conditions with its value for each of the 5 parameters. I mean , first I need to know how to create f_x.dry

Answer 1

[dpb]

If you have a recent release, put the coefficient data into a table object...

http://www.mathworks.com/help/matlab/matlab_prog/access-data-in-a-table.html

If not, use a named structure or a named index variable into an array. Syntax wouldn't be quite so clean but would accomplish the deed.

[Priya]

Thanks. I will try this out and will let you know.

[Priya]

Could you please find the above code and help me how to implement this.

Answer 2

[dpb]

Well, I don't have the table object here so can't; but without it the outline of what I'd do is sotoo...

%Parameters   Dry  wet  Low  very low
k_A=[1.00 1.00 1.00 1.00];
k_S=[0.40 0.40 0.40 0.40];
cof=[0.55 0.30 0.06 0.03];  
A=[0.40 0.40 0.40 0.40];  
B=[0.60 0.20 0.20 0.10];  

V_wx=40;
V_rx=20;
V_wy=21;
V_ry=20;
V_f=(V_wx+V_rx)/2;
s_x=(V_wx-V_rx)/V_f;
s_y=(V_wy-V_ry)/V_f;
s=sqrt(s_x^2+s_y^2);    
w_x=V_wx-V_rx;          
w_y=V_wy-V_ry;  
w=w_x+w_y; 

c=@(ix) cof(ix).*((1-A(ix)).*exp(-B(ix).*w)+A(ix));

ix=1:length(k_A); ix=ix(:);
[ix c(ix).']

Here I've revamped your data tables to shorten them to the base names with the intent of using a value directly from the array instead of thru some other intermediary variable ("the Matlab way"). Then I recast the initial calculation of the coefficient variable cof to be an anonymous function c that can be invoked with either a single value or a vector of values of the various pieces that make it up.

Lastly, to demonstrate its use, I set the index vector to the possible choices and invoked the function thru the variable containing it and displayed the input index and the result for each at the command line.

As far as the desire for a name associated with the condition that could be used, you could do something like

params={'Dry' 'Wet' 'Low' 'Very Low'}.';
ix=listdlg('PromptString','Select a condition:',...
           'SelectionMode','single',...
           'ListString',params)

at the beginning to select a particular index w/ user interaction.

From here it shouldn't take too much to recast the rest of your code in a similar vein.