Ant colony optimization algorithms
In computer science and operations research, the ant colony optimization algorithm is a probabilistic technique for solving computational problems which can be reduced to finding good paths through graphs. Artificial Ants stand for multi-agent methods inspired by the behavior of real ants.
The pheromone-based communication of biological ants is often the predominant paradigm used. Combinations of Artificial Ants and local search algorithms have become a method of choice for numerous optimization tasks involving some sort of graph, e.g., vehicle routing and internet routing. The burgeoning activity in this field has led to conferences dedicated solely to Artificial Ants, and to numerous commercial applications by specialized companies such as AntOptima.
As an example, Ant colony optimization is a class of optimization algorithms modeled on the actions of an ant colony. Artificial 'ants' locate optimal solutions by moving through a parameter space representing all possible solutions. Real ants lay down pheromones directing each other to resources while exploring their environment. The simulated 'ants' similarly record their positions and the quality of their solutions, so that in later simulation iterations more ants locate better solutions. One variation on this approach is the bees algorithm, which is more analogous to the foraging patterns of the honey bee, another social insect.
This algorithm is a member of the ant colony algorithms family, in swarm intelligence methods, and it constitutes some metaheuristic optimizations. Initially proposed by Marco Dorigo in 1992 in his PhD thesis, the first algorithm was aiming to search for an optimal path in a graph, based on the behavior of ants seeking a path between their colony and a source of food. The original idea has since diversified to solve a wider class of numerical problems, and as a result, several problems have emerged, drawing on various aspects of the behavior of ants. From a broader perspective, ACO performs a model-based search and shares some similarities with estimation of distribution algorithms.
Overview
In the natural world, ants of some species wander randomly, and upon finding food return to their colony while laying down pheromone trails. If other ants find such a path, they are likely not to keep travelling at random, but instead to follow the trail, returning and reinforcing it if they eventually find food.Over time, however, the pheromone trail starts to evaporate, thus reducing its attractive strength. The more time it takes for an ant to travel down the path and back again, the more time the pheromones have to evaporate. A short path, by comparison, gets marched over more frequently, and thus the pheromone density becomes higher on shorter paths than longer ones. Pheromone evaporation also has the advantage of avoiding the convergence to a locally optimal solution. If there were no evaporation at all, the paths chosen by the first ants would tend to be excessively attractive to the following ones. In that case, the exploration of the solution space would be constrained. The influence of pheromone evaporation in real ant systems is unclear, but it is very important in artificial systems.
The overall result is that when one ant finds a good path from the colony to a food source, other ants are more likely to follow that path, and positive feedback eventually leads to many ants following a single path. The idea of the ant colony algorithm is to mimic this behavior with "simulated ants" walking around the graph representing the problem to solve.
Ambient networks of intelligent objects
New concepts are required since “intelligence” is no longer centralized but can be found throughout all minuscule objects. Anthropocentric concepts have been known to lead to the production of IT systems in which data processing, control units and calculating forces are centralized. These centralized units have continually increased their performance and can be compared to the human brain. The model of the brain has become the ultimate vision of computers. Ambient networks of intelligent objects and, sooner or later, a new generation of information systems which are even more diffused and based on nanotechnology, will profoundly change this concept. Small devices that can be compared to insects do not dispose of a high intelligence on their own. Indeed, their intelligence can be classed as fairly limited. It is, for example, impossible to integrate a high performance calculator with the power to solve any kind of mathematical problem into a biochip that is implanted into the human body or integrated in an intelligent tag which is designed to trace commercial articles. However, once those objects are interconnected they dispose of a form of intelligence that can be compared to a colony of ants or bees. In the case of certain problems, this type of intelligence can be superior to the reasoning of a centralized system similar to the brain.Nature offers several examples of how minuscule organisms, if they all follow the same basic rule, can create a form of collective intelligence on the macroscopic level. Colonies of social insects perfectly illustrate this model which greatly differs from human societies. This model is based on the co-operation of independent units with simple and unpredictable behavior. They move through their surrounding area to carry out certain tasks and only possess a very limited amount of information to do so. A colony of ants, for example, represents numerous qualities that can also be applied to a network of ambient objects. Colonies of ants have a very high capacity to adapt themselves to changes in the environment as well as an enormous strength in dealing with situations where one individual fails to carry out a given task. This kind of flexibility would also be very useful for mobile networks of objects which are perpetually developing. Parcels of information that move from a computer to a digital object behave in the same way as ants would do. They move through the network and pass from one knot to the next with the objective of arriving at their final destination as quickly as possible.
Artificial pheromone system
Pheromone-based communication is one of the most effective ways of communication which is widely observed in nature. Pheromone is used by social insects such asbees, ants and termites; both for inter-agent and agent-swarm communications. Due to its feasibility, artificial pheromones have been adopted in multi-robot and swarm robotic systems. Pheromone-based communication was implemented by different means such as chemical or physical ways. However, those implementations were not able to replicate all the aspects of pheromones as seen in nature.
Using projected light was presented in an 2007 IEEE paper by Garnier, Simon, et al. as an experimental setup to study pheromone-based communication with micro autonomous robots. Another study that proposed a novel pheromone communication method, COSΦ, for a swarm robotic system is based on precise and fast visual localization.
The system allows simulation of a virtually unlimited number of different pheromones and provides the result of their interaction as a gray-scale image on a horizontal LCD screen that the robots move on. In order to demonstrate the pheromone communication method, Colias autonomous micro robot was deployed as the swarm robotic platform.
Algorithm and formulae
In the ant colony optimization algorithms, an artificial ant is a simple computational agent that searches for good solutions to a given optimization problem. To apply an ant colony algorithm, the optimization problem needs to be converted into the problem of finding the shortest path on a weighted graph. In the first step of each iteration, each ant stochastically constructs a solution, i.e. the order in which the edges in the graph should be followed. In the second step, the paths found by the different ants are compared. The last step consists of updating the pheromone levels on each edge.procedure ACO_MetaHeuristic is
while not_termination do
generateSolutions
daemonActions
pheromoneUpdate
repeat
end procedure
Edge selection
Each ant needs to construct a solution to move through the graph. To select the next edge in its tour, an ant will consider the length of each edge available from its current position, as well as the corresponding pheromone level. At each step of the algorithm, each ant moves from a state to state, corresponding to a more complete intermediate solution. Thus, each ant computes a set of feasible expansions to its current state in each iteration, and moves to one of these in probability. For ant, the probability of moving from state to state depends on the combination of two values, the attractiveness of the move, as computed by some heuristic indicating the a priori desirability of that move and the trail level of the move, indicating how proficient it has been in the past to make that particular move. The trail level represents a posteriori indication of the desirability of that move.In general, the th ant moves from state to state with probability
where
is the amount of pheromone deposited for transition from state to, 0 ≤ is a parameter to control the influence of, is the desirability of state transition and ≥ 1 is a parameter to control the influence of. and represent the trail level and attractiveness for the other possible state transitions.
Pheromone update
Trails are usually updated when all ants have completed their solution, increasing or decreasing the level of trails corresponding to moves that were part of "good" or "bad" solutions, respectively. An example of a global pheromone updating rule iswhere is the amount of pheromone deposited for a state transition, is the pheromone evaporation coefficient and is the amount of pheromone deposited by th ant, typically given for a TSP problem by
where is the cost of the th ant's tour and is a constant.
Common extensions
Here are some of the most popular variations of ACO algorithms.Ant System (AS)
The Ant System is the first ACO algorithm. This algorithm corresponds to the one presented above. It was developed by Dorigo.Ant Colony System (ACS)
In the Ant Colony System algorithm, the original Ant System was modified in three aspects: the edge selection is biased towards exploitation ; while building a solution, ants change the pheromone level of the edges they are selecting by applying a local pheromone updating rule; at the end of each iteration, only the best ant is allowed to update the trails by applying a modified global pheromone updating rule.Elitist Ant System
In this algorithm, the global best solution deposits pheromone on its trail after every iteration, along with all the other ants.Max-Min Ant System (MMAS)
This algorithm controls the maximum and minimum pheromone amounts on each trail. Only the global best tour or the iteration best tour are allowed to add pheromone to its trail. To avoid stagnation of the search algorithm, the range of possible pheromone amounts on each trail is limited to an interval . All edges are initialized to τmax to force a higher exploration of solutions. The trails are reinitialized to τmax when nearing stagnation.Rank-based Ant System (ASrank)
All solutions are ranked according to their length. Only a fixed number of the best ants in this iteration are allowed to update their trials. The amount of pheromone deposited is weighted for each solution, such that solutions with shorter paths deposit more pheromone than the solutions with longer paths.Continuous Orthogonal Ant Colony (COAC)
The pheromone deposit mechanism of COAC is to enable ants to search for solutions collaboratively and effectively. By using an orthogonal design method, ants in the feasible domain can explore their chosen regions rapidly and efficiently, with enhanced global search capability and accuracy. The orthogonal design method and the adaptive radius adjustment method can also be extended to other optimization algorithms for delivering wider advantages in solving practical problems.Recursive Ant Colony Optimization
It is a recursive form of ant system which divides the whole search domain into several sub-domains and solves the objective on these subdomains. The results from all the subdomains are compared and the best few of them are promoted for the next level. The subdomains corresponding to the selected results are further subdivided and the process is repeated until an output of desired precision is obtained. This method has been tested on ill-posed geophysical inversion problems and works well.Convergence
For some versions of the algorithm, it is possible to prove that it is convergent. The first evidence of convergence for an ant colony algorithm was made in 2000, the graph-based ant system algorithm, and later on for the ACS and MMAS algorithms. Like most metaheuristics, it is very difficult to estimate the theoretical speed of convergence. A performance analysis of a continuous ant colony algorithm with respect to its various parameters showed that its performance and rate of convergence are sensitive to the chosen parameter values, and especially to the value of the pheromone evaporation rate. In 2004, Zlochin and his colleagues showed that COAC-type algorithms could be assimilated methods of stochastic gradient descent, on the cross-entropy and estimation of distribution algorithm. They proposed these metaheuristics as a "research-based model".Applications
Ant colony optimization algorithms have been applied to many combinatorial optimization problems, ranging from quadratic assignment to protein folding or routing vehicles and a lot of derived methods have been adapted to dynamic problems in real variables, stochastic problems, multi-targets and parallel implementations.It has also been used to produce near-optimal solutions to the travelling salesman problem. They have an advantage over simulated annealing and genetic algorithm approaches of similar problems when the graph may change dynamically; the ant colony algorithm can be run continuously and adapt to changes in real time. This is of interest in network routing and urban transportation systems.
The first ACO algorithm was called the ant system and it was aimed to solve the travelling salesman problem, in which the goal is to find the shortest round-trip to link a series of cities. The general algorithm is relatively simple and based on a set of ants, each making one of the possible round-trips along the cities. At each stage, the ant chooses to move from one city to another according to some rules:
- It must visit each city exactly once;
- A distant city has less chance of being chosen ;
- The more intense the pheromone trail laid out on an edge between two cities, the greater the probability that that edge will be chosen;
- Having completed its journey, the ant deposits more pheromones on all edges it traversed, if the journey is short;
- After each iteration, trails of pheromones evaporate.
Scheduling problem
- Sequential Ordering Problem
- Job-shop scheduling problem
- Open-shop scheduling problem
- Permutation flow shop problem
- Single machine total tardiness problem
- Single machine total weighted tardiness problem
- Resource-constrained project scheduling problem
- Group-shop scheduling problem
- Single-machine total tardiness problem with sequence dependent setup times
- Multistage flowshop scheduling problem with sequence dependent setup/changeover times
Vehicle routing problem
- Capacitated vehicle routing problem
- Multi-depot vehicle routing problem
- Period vehicle routing problem
- Split delivery vehicle routing problem
- Stochastic vehicle routing problem
- Vehicle routing problem with pick-up and delivery
- Vehicle routing problem with time windows
- Time dependent vehicle routing problem with time windows
- Vehicle routing problem with time windows and multiple service workers
Assignment problem
- Quadratic assignment problem
- Generalized assignment problem
- Frequency assignment problem
- Redundancy allocation problem
Set problem
- Set cover problem
- Partition problem
- Weight constrained graph tree partition problem
- Arc-weighted l-cardinality tree problem
- Multiple knapsack problem
- Maximum independent set problem
Device sizing problem in nanoelectronics physical design
- Ant colony optimization based optimization of 45 nm CMOS-based sense amplifier circuit could converge to optimal solutions in very minimal time.
- Ant colony optimization based reversible circuit synthesis could improve efficiency significantly.
Antennas optimization and synthesis
Image processing
The ACO algorithm is used in image processing for image edge detection and edge linking.- Edge detection:
The following are the steps involved in edge detection using ACO:
Step1: Initialization:
Randomly place ants on the image where . Pheromone matrix are initialized with a random value. The major challenge in the initialization process is determining the heuristic matrix.
There are various methods to determine the heuristic matrix. For the below example the heuristic matrix was calculated based on the local statistics:
the local statistics at the pixel position.
Where is the image of size
, which is a normalization factor
can be calculated using the following functions:
The parameter in each of above functions adjusts the functions’ respective shapes.
Step 2 Construction process:
The ant's movement is based on 4-connected pixels or 8-connected pixels. The probability with which the ant moves is given by the probability equation
Step 3 and Step 5 Update process:
The pheromone matrix is updated twice. in step 3 the trail of the ant is updated where as in step 5 the evaporation rate of the trail is updated which is given by the below equation.
, where is the pheromone decay coefficient
Step 7 Decision Process:
Once the K ants have moved a fixed distance L for N iteration, the decision whether it is an edge or not is based on the threshold T on the pheromone matrixτ. Threshold for the below example is calculated based on Otsu's method.
Image Edge detected using ACO:
The images below are generated using different functions given by the equation to.
- Edge linking: ACO has also been proven effective in edge linking algorithms too.
Other applications
- Bankruptcy prediction
- Classification
- Connection-oriented network routing
- Connectionless network routing
- Data mining
- Discounted cash flows in project scheduling
- Distributed information retrieval
- Energy and electricity network design
- Grid workflow scheduling problem
- Inhibitory peptide design for protein protein interactions
- Intelligent testing system
- Power electronic circuit design
- Protein folding
- System identification
Definition difficulty
Stigmergy algorithms
There is in practice a large number of algorithms claiming to be "ant colonies", without always sharing the general framework of optimization by canonical ant colonies. In practice, the use of an exchange of information between ants via the environment is deemed enough for an algorithm to belong to the class of ant colony algorithms. This principle has led some authors to create the term "value" to organize methods and behavior based on search of food, sorting larvae, division of labour and cooperative transportation.Related methods
- Genetic algorithms maintain a pool of solutions rather than just one. The process of finding superior solutions mimics that of evolution, with solutions being combined or mutated to alter the pool of solutions, with solutions of inferior quality being discarded.
- An estimation of distribution algorithm is an evolutionary algorithm that substitutes traditional reproduction operators by model-guided operators. Such models are learned from the population by employing machine learning techniques and represented as probabilistic graphical models, from which new solutions can be sampled or generated from guided-crossover.
- Simulated annealing is a related global optimization technique which traverses the search space by generating neighboring solutions of the current solution. A superior neighbor is always accepted. An inferior neighbor is accepted probabilistically based on the difference in quality and a temperature parameter. The temperature parameter is modified as the algorithm progresses to alter the nature of the search.
- Reactive search optimization focuses on combining machine learning with optimization, by adding an internal feedback loop to self-tune the free parameters of an algorithm to the characteristics of the problem, of the instance, and of the local situation around the current solution.
- Tabu search is similar to simulated annealing in that both traverse the solution space by testing mutations of an individual solution. While simulated annealing generates only one mutated solution, tabu search generates many mutated solutions and moves to the solution with the lowest fitness of those generated. To prevent cycling and encourage greater movement through the solution space, a tabu list is maintained of partial or complete solutions. It is forbidden to move to a solution that contains elements of the tabu list, which is updated as the solution traverses the solution space.
- Artificial immune system algorithms are modeled on vertebrate immune systems.
- Particle swarm optimization, a swarm intelligence method
- Intelligent water drops, a swarm-based optimization algorithm based on natural water drops flowing in rivers
- Gravitational search algorithm, a swarm intelligence method
- Ant colony clustering method, a method that make use of clustering approach, extending the ACO.
- Stochastic diffusion search, an agent-based probabilistic global search and optimization technique best suited to problems where the objective function can be decomposed into multiple independent partial-functions
History
Chronology of ant colony optimization algorithms.
- 1959, Pierre-Paul Grassé invented the theory of stigmergy to explain the behavior of nest building in termites;
- 1983, Deneubourg and his colleagues studied the collective behavior of ants;
- 1988, and Moyson Manderick have an article on self-organization among ants;
- 1989, the work of Goss, Aron, Deneubourg and Pasteels on the collective behavior of Argentine ants, which will give the idea of ant colony optimization algorithms;
- 1989, implementation of a model of behavior for food by Ebling and his colleagues;
- 1991, M. Dorigo proposed the ant system in his doctoral thesis. A technical report extracted from the thesis and co-authored by V. Maniezzo and A. Colorni was published five years later;
- 1994, Appleby and Steward of British Telecommunications Plc published the first application to telecommunications networks
- 1995, Gambardella and Dorigo proposed ant-q, the preliminary version of ant colony system as first estension of ant system;.
- 1996, Gambardella and Dorigo proposed ant colony system
- 1996, publication of the article on ant system;
- 1996, Hoos and Stützle invent the max-min ant system;
- 1997, Dorigo and Gambardella proposed ant colony system hybridized with local search;
- 1997, Schoonderwoerd and his colleagues published an improved application to telecommunication networks;
- 1998, Dorigo launches first conference dedicated to the ACO algorithms;
- 1998, Stützle proposes initial parallel implementations;
- 1999, Gambardella, Taillard and Agazzi proposed macs-vrptw, first multi ant colony system applied to vehicle routing problems with time windows,
- 1999, Bonabeau, Dorigo and Theraulaz publish a book dealing mainly with artificial ants
- 2000, special issue of the Future Generation Computer Systems journal on ant algorithms
- 2000, first applications to the scheduling, scheduling sequence and the satisfaction of constraints;
- 2000, Gutjahr provides the first evidence of convergence for an algorithm of ant colonies
- 2001, the first use of COA algorithms by companies ;
- 2001, Iredi and his colleagues published the first multi-objective algorithm
- 2002, first applications in the design of schedule, Bayesian networks;
- 2002, Bianchi and her colleagues suggested the first algorithm for stochastic problem;
- 2004, Dorigo and Stützle publish the Ant Colony Optimization book with MIT Press
- 2004, Zlochin and Dorigo show that some algorithms are equivalent to the stochastic gradient descent, the cross-entropy method and algorithms to estimate distribution
- 2005, first applications to protein folding problems.
- 2012, Prabhakar and colleagues publish research relating to the operation of individual ants communicating in tandem without pheromones, mirroring the principles of computer network organization. The communication model has been compared to the Transmission Control Protocol.
- 2016, first application to peptide sequence design.
- 2017, successful integration of the multi-criteria decision-making method PROMETHEE into the ACO algorithm.
Publications (selected)
- M. Dorigo, 1992. Optimization, Learning and Natural Algorithms, PhD thesis, Politecnico di Milano, Italy.
- M. Dorigo, V. Maniezzo & A. Colorni, 1996. "", IEEE Transactions on Systems, Man, and Cybernetics–Part B, 26 : 29–41.
- M. Dorigo & L. M. Gambardella, 1997. "". IEEE Transactions on Evolutionary Computation, 1 : 53–66.
- M. Dorigo, G. Di Caro & L. M. Gambardella, 1999. "". Artificial Life, 5 : 137–172.
- E. Bonabeau, M. Dorigo et G. Theraulaz, 1999. Swarm Intelligence: From Natural to Artificial Systems, Oxford University Press.
- M. Dorigo & T. Stützle, 2004. Ant Colony Optimization, MIT Press.
- M. Dorigo, 2007. . Scholarpedia.
- C. Blum, 2005 "". Physics of Life Reviews, 2: 353-373
- M. Dorigo, M. Birattari & T. Stützle, 2006 '. TR/IRIDIA/2006-023
- Mohd Murtadha Mohamad,"Articulated Robots Motion Planning Using Foraging Ant Strategy",Journal of Information Technology - Special Issues in Artificial Intelligence, Vol.20, No. 4 pp. 163–181, December 2008,.
- N. Monmarché, F. Guinand & P. Siarry, "Artificial Ants", August 2010 Hardback 576 pp. .
- A. Kazharov, V. Kureichik, 2010. "", Journal of Computer and Systems Sciences International, Vol. 49. No. 1. pp. 30–43.
- C-M. Pintea, 2014, , Springer
- K. Saleem, N. Fisal, M. A. Baharudin, A. A. Ahmed, S. Hafizah and S. Kamilah, "Ant colony inspired self-optimized routing protocol based on cross layer architecture for wireless sensor networks", WSEAS Trans. Commun., vol. 9, no. 10, pp. 669–678, 2010.
- K. Saleem and N. Fisal, "Enhanced Ant Colony algorithm for self-optimized data assured routing in wireless sensor networks", Networks 2012 18th IEEE International Conference on, pp. 422–427.
- Abolmaali S, Roodposhti FR. ' Journal of Accounting. 2018 Mar;8.