7/31/2019 04_Comparison of Met a Heuristic Algorithms

1/13

4. Comparison of Metaheuristic

In the name of GodIn the name of God

Comparison of Metaheuristic Algorithms

Fall 2010

Instructor: Dr. Masoud Yaghini

7/31/2019 04_Comparison of Met a Heuristic Algorithms

2/13

Emergence Time

GA: 1975

ACO: 1991

TS: 1986

SA: 1986

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

3/13

Innovator

GA: John Holland

ACO: Marco Dorigo

TS: Fred Glover

SA: A. Kirkpatrick

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

4/13

Source of Inspiration

GA: Evolution Principle

ACO: The foraging behavior of ant colonies

TS: a prohibition imposed by social custom as aprotective measure or something banned as

constituting a risk.

Comparison of Metaheuristic Algorithms

SA: physical annealing

7/31/2019 04_Comparison of Met a Heuristic Algorithms

5/13

Originally Purpose

GA: for solving combinatorial problems

ACO: for solving combinatorial problems

TS: for solving combinatorial problems

SA: for solving combinatorial problems

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

6/13

Population or Single solution orientation

GA: Population-based algorithm

ACO: Population & Single based

TS: Single based

SA: Single solution

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

7/13

Using Memory

GA: Memory less

ACO: Using memory to store amount of pheromones

TS: Short term (tabu lists), midterm and long termmemory

SA: Memor less

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

8/13

Generating Initial Solution

GA: Random

ACO: Random / Local search

TS: Local search

SA: Random

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

9/13

Finding Neighbor Solutions

GA: Random search of neighbouring solutions

using mutation operator

ACO: Random search of neighbouring solutions using random proportional rule

TS: Deterministic/random search of nei hbourin

Comparison of Metaheuristic Algorithms

solutions

SA: Random search of neighbouring solutions

7/31/2019 04_Comparison of Met a Heuristic Algorithms

10/13

Number of Neighbour Solutions Considered at Each Move

GA: One neighbour solution

ACO: n neighbour solutions

TS: n neighbour solutions

SA: One neighbour solution

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

11/13

Finding Local Optimum

GA: Mutation operator

ACO: Accumulation pheromone on better solutions

TS: Based on a local search

SA: Decreasing temperature and limiting search space

Comparison of Metaheuristic Algorithms

7/31/2019 04_Comparison of Met a Heuristic Algorithms

12/13

Escaping from Local Optimum

GA: Random search of search space

Using crossover operator

ACO: Evaporation mechanism TS: Deterministic acceptance of non-improving

solutions & jump to unvisited search space

Comparison of Metaheuristic Algorithms

Using tabu lists and diversification

SA: Probabilistic acceptance of non-improving

solutions

based on acceptance function and temperature parameter

7/31/2019 04_Comparison of Met a Heuristic Algorithms

13/13

Comparison of Metaheuristic Algorithms