4
$\begingroup$

I'm learning fuzzy logic and more or less understand the basic concept, but i'm having a hard time understanding how to apply it to a method. I tried browsing online for explanation on how to use it, but only found some implementation and test case using the basic form of 4 rules and 3 variables, and 2 rules per variable. Anyway this is an example case, i will use Tsukamoto method.

In this case actually i have 6 rules and 3 variables with 3 rules ver variable, but i will only explain 1 of the variables because i think the rest will have the same solution. I have 3 variables one of them is "size", the range is for small it's 0-2 and for large it's 7-10. The current condition is size = 6.5. The rules is as follow(simplified to only use this variable):

  • [R1] size = small
  • [R2] size = medium
  • [R3] size = large

What i want to know is:

  • how do i define the formula for medium(the middle rule if the case is different)?
  • what if the rule is more than 3 (i.e. small, medium, large, ex-large)?

What i understands if the rule is only 2 i can use this formula

  • small[x]=(max-x)/(max-min)
  • large[x]=(x-min)/(max-min)

My current approach to this problem is as follow:

small[x]=1; x<=2

medium[x]=(max-x)/(max-min); 2 < x < 7

large[x]=0; x>=7

Is this correct? Also can you refer me to some source to study this? As i mentioned before i can only find some implementation and basic explanation, it's either there is no online source for this or i don't know what to search for. Sorry if it's hard to understand i can edit and post the whole problem if you want, thanks in advance.

Extra question: what is the name of algorithm which can be used to solve the crossing bridge puzzle(the one with timer, max person, and stuff)?i forgot the name.

$\endgroup$
  • 1
    $\begingroup$ Is the question about understanding the Tsukamoto method? The method applies to n multivariate rules, but first, it is not clear how you applied the method here. $\endgroup$ – Eric Platon Jun 4 '18 at 14:29
  • 1
    $\begingroup$ No it's the membership graph and function that I don't understand. I am confused at what function/formula should i use and what values should i add to the function if the value of a variable is a range in above example large = 7-10. I know a bit to proceed after that, only the basic though but I want to understand this one first. $\endgroup$ – Dizz Jun 4 '18 at 15:27
0
$\begingroup$

After i did a bit more study on this, i found this solution.

Membership Graph

μsmall[x]

  • 1; -> x≤0

  • (20-x)/(20-0); -> 0

  • 0; -> x≥20

μmedium[x]

  • 0; -> x≤20 or x≥70

  • (x-20)/(45-20); -> 20

  • (70-x)/(70-45); -> 45

μlarge[x]

  • 0; -> x≤70
  • (x-70)/(100-70); -> 70
  • 1; -> x≥100

For the final result i need to add 3 more rules for the function to work. So it become 9 rules and 3 variables, with 3 rules for each input variables.

To answer both of my question:

  1. Add a triangle graph and use the formula.

  2. Add a triangle graph for each rules.

What i learn from this is the total rules needed to determine the output is the number of input variables rules multiplied(i.e. v1 = 3 rules, v2 3 rules, total rules = 3x3=9 rules). And i'm not really sure about this one but you can combine multiple kind of graphs.

$\endgroup$
  • $\begingroup$ I know this is a bad practice because it's basically asking clarification for my answer. But because i'm still new at this and not really sure about my answer, i will mark my answer as accepted after someone can verify it. $\endgroup$ – Dizz Jun 6 '18 at 7:36
  • $\begingroup$ Just a philosophical add, fuzzy logic is an application of the grey between absolutes, and absolute are associated a quantity to define there extremes , But the grey between is 'uncertainty' but the application uses quantity to measure the degree between the extremes and I don't think the grey of uncertainty should be given an absolute measure of degree, as i don't think the philosophy it sprang from would 'measure' the nature of its 'uncertainty' they way that mathematics does. $\endgroup$ – Bobs Jun 7 '18 at 21:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.