วันอังคารที่ 19 สิงหาคม พ.ศ. 2557

After noon class with Computer Intelligence by Pro.Chu part 1

This story is more about mathematics so a bit  hard but keep try it's not too hard so
Starting with simple OR stand for Operation Research
What can we make the thing

Optimize under constraints condition 

Heuristics : exact the problem and find the exact solution by using algorithm
example : Simulated Annealing for physic

Metaheuristics: nearly optimize solution

Random search waste time and not guarantee the solution

next is airline hub and floor layout what is make sense to be use as a hub by planing or layout under condition time and less energy and other condition if in floor. size is matter for this condition

Critical terms
Optimize Problem:
What is Objective function and Quality that call  fitness.
F(X1 + X2 + X3 )= X1*P2+  X2*P2 +X3*P3
F = FITNESS VALUE 
X1, X2, X3 = DECISION VARIABLE
search space 
constraints = ข้อจำกัด
boundary = ขอบเขต

min f(x,y) where f(x,y) = x^2 + y^2 , where x,y E [-10,10]

D = Dimension , Dimenallity  = number of variable

The number of variables
- univariate problem = 1 var
- multivariate problem = 2 or more

Degree of nonlinearity of the obj function:
 - Linear = f=x+y
 - Quadratic = (x+y)^2  this is power 2
 -  Nonlinear = other power that not 2 is all non linear

Type of variables 
 - Continuous problem ex: xi E R
 - Discrete or integer opt problem ex: xi E T
 - Mixed opt problem
- Combinatorial opt problem = swap the way of point ar like shuffle the way to make the opt way that fix problem the best other like is traveling salesman, vehicle routing problem or nurse/ guard schedule.

More constrain problem
 equality and inequality Feasible region like X1+X2+X3 <= 10 OR 3X1 + X3 <=24
more condition more boundary constraint that make search space or feasible region is less more that good


MySimpleDiary
Continue to part 2

ไม่มีความคิดเห็น:

แสดงความคิดเห็น