Information Technology Reference
In-Depth Information
Evolutionary programming is capable of reproducing new computer programs
so as to solve problems. It consists of three steps:
1. Generate initial population, which randomly consists of the functions(programs)
related to the problem.
2. Repeat the following sub-steps, until selection criterion is satisfied:
①
Perform each program in population and endow it with a fitness value
according to its ability to solve problem.
②
Use related operations, such as mutation, to generate new computer
programs, and then form a new population. Probabilistically select a
computer program from population based on fitness value and then use a
related operation to deal with this program. Replicate existing computer
program to a new population. Create new computer programs by
randomly recombining two computer programs
3.
In the final population, the computer program with the highest fitness value
is regarded as the result of evolutionary programming. This result is possibly
the optimal or approximate solution to the problem.
Exercises
1. Describe the theoretical model of evolution systems.
2. Give Darwin's evolutionary algorithm.
3. Expatiate the definition of classical genetic algorithm and explain its basic
elements.
4. Genetic algorithm attempts to keep genetic diversity as well as some important
features(genetic schemas). Please design a different genetic operator, so that
genetic algorithm can reach the both targets.
5. What is simple mutation? Please give its brief description.
6. What is simple inversion operation? Please give its brief description.
7. For TSP problem, discuss the problem of its effective coding and design
related genetic operations and fitness function.
8. Design a genetic algorithm for solving TSP problem.
9. Please explain how to express chromosomes and genes in genetic algorithm.
10. For two parent individuals 1110###0# and 1##0111##, suppose that crossover
location is 6. Please use alone point crossing operator to create their
offspring.
11. Please use a programming language to develop a program for parallel
genetic algorithm.
