Question

Can anyone please sort the differential evolution alogrithim for the sake of automatic time table generator?

Answer 1

Differnetial Evaluation Algorithm for the sake of automatic time table generator :-

The strength of differential evolution’s approach is that it display better results than a genetic algorithm and other evolutionary algorithms and can be easily applied to a wide variety of real valued problems despite noisy,multi-dimensional spaces, which usually make the problems very difficult for optimization. Another impressive trait of differential evolution is that the parameters CR and F do not require the same fine tuning which is necessary in many other evolutionary algorithms. Differential evolution has been used effectively in many applications on various domains such as neural network learning, image processing etc.

The differential evolution algo. can be applied to real-valued problems over a continuous space with much more ease than a genetic algorithm. The idea is that the difference between two vectors yields a difference vector which can be used with a scaling factor to traverse the search space. As in the genetic algorithm’s beginning a random population is chosen, equally over the problem space, and to create the next generation an equal number of donor vectors are created through means of

i n : Di = Xa + F(Xb − Xc) where i, a, b, c are distinct .

where Xb and Xc are randomly chosen and Xa is chosen either randomly or as one of the best members of the population . A trial vector Ti, j is created by choosing between the donor vector and the previous generation for each element (j) according to the crossover rate CR[0–1], for each element in the vector we choose either the corresponding element from the previous generation vector or from the donor vector such that

i, j : if (random < CRkj = Jrand) then Ti,j = Di,j otherwise Ti,j = Xi,j

where Jr and is randomly chosen for each iteration through i and ensures that no Ti is exactly the same as the corresponding Xi. Then the trial vector’s fitness is evaluated, and for each member of the new generation, X′i , we choose the better performing of the previous generation, Xi, or the trial vector, Ti.

algo:-

Begin

generate randomly an initial population of solutions.

Calculate the fitness of the initial population.

Now Repeat,

For each parent, select three solutions at random.

Create one offspring using the DE operators.

Do this a number of times equal to the population size.

for each member of the next generation

If offspring (x) is more fit then parent(x), Parent (x) is replaced.

Untill a stop condition is satisfied.

End.