This repository provides the implementations of evalution algorithm introduced in Sustech-CS408.
The main task of this section is applying Evolutionary Algorithms (EAs) to optimising Travelling Salesman Problems (TSPs).
The traveling salesman problem(TSP) is a combinatorial optimization problem that can be mathematically modeled as a problem of searching the optimal permutation.
This research describes a model for mutation in the evolution algorithm(EA) and adopts group theory to the demonstration of the global optimal promise for mutation.
Also, this research uses a simple framework of EA to compare the performances of different mutation operators.
The experimental results show the degeneration under this simple framework and all the mutation operators perform the same well as each other.
The Evolutionary Algorithms (EAs) are inspired by nature.
EAs designed with a niching mechanism can locate multiple globally optimal or suboptimal solutions for the problem of optimization.
In this work, a simple EA with crowding and fitness sharing is designed for optimizing multimodal functions given by Benchmark for CEC’2013 special session and competition.
Fitness sharing is a double-edged sword: On one hand, it drives the divergent evolution; on the other hand, it deceives the agents or the selector, making them mistaken for the global optimal. But actually, they are not, not even the local optimal. This kind of illusion affects their ability to locate global optimals.
The capacitated arc routing problem (CARP) is a classical challenging combinatorial optimization problem with wide practice applications.
Many metaheuristic methods are applied to solve CARP.
In this report, a similar algorithm with the method described in A Scalable Approach to Capacitated Arc Routing Problems Based on Hierarchical Decomposition (SAHiD) is proposed.
The experimental results show that there is still a lot of room for improvement in this algorithm.