Oblicz. miękkie-metody i zast.

Kod programu na którym opiera się projekt.

Grupa zajęciowa

Rok akademicki 2014/2015

Semestr zimowy

Wt. 9:15

Autorzy

• Jacek Miszczak (179158)
• Filip Malczak (179326)

Following text should not be graded or taken under any consideration while grading. These are only developer notes.

Dev Notes

Following proposals are implemented in this version. Feel free to comment and not agree. Proposal: comments as lists under each proposal; it would (hopefully) leave nice thought stack.

Proposal: Where possible, use list literals (ArrayLists are the fastest non-array linear data structure). ~F

Proposal: Represent evaluations as double and probability as integer (attribute of value x would mean probability of x/1000) ~F

Note: Components using Evolution Context (see dictionary for all capitalized terms) are not forced to work together by any type-safety techniques. You can create ContextManager that will not provide informations needed by your implementation of CP, and that is your fault. The contract makes keeping it consistent your duty. ~F

Dictionary

• Specimen - object in searched space (function tree i some form, in case of our problem)
• Population - list of Specimen s
• Evaluation - function from searched space to real numbers (represented as double), which should be minimised
• Generation - number of processed population; number of genetic algorithms iteration
• CP - crossover probability; may change with each generation
• MP - mutation probability; may change with each generation
• (Evolution) Context - namespace to keep all statistics on past and current populations, for parameter adaptation sake
• Context handler - piece of logic that takes care of filling context with statistics at the beginning of each GA iteration.