The concept of ant colony optimization (ACO) is based on the efficient and effective way that ant colonies find food. At its heart, this behaviour is the ants' use of chemical pheromone trails for indirect communication with one another, which allows them to locate efficient routes between their nest and food sources. 

"In the literal sense, the programmed computer understands what the car or the adding machine understand: namely, exactly nothing." - John Searle, 1980

Both the scientific and industrial communities rely heavily on solutions to optimization challenges. For example, timetable optimization, nurse shift scheduling, train scheduling, capacity planning, vehicle routing difficulties, group shop scheduling, portfolio optimization, etc., are all real-world applications of optimization theory and practice. It is why a lot of effort goes into creating optimization algorithms. 

Ant colony optimization

Probabilistic ant colony optimization seeks out the best actions to take. The ant colony optimization algorithm solves various computational problems in computer science and related fields. For example, ACO is a popular optimization algorithm that uses the probabilistic technique for computational problem-solving and optimal path-finding with graphs.

Artificial 'ants' (e.g., simulation agents) explore a parameter space representing all possible solutions to find the optimal one. For example, while foraging for food, real ants mark their territory with pheromones that guide other ants to nearby food sources. The simulated 'ants' also keep track of their locations and the quality of the solutions they find, allowing for more ants to find better answers in subsequent simulation cycles. The bees algorithm is a variant on this theme that takes cues from the foraging behaviour of honey bees.

Metaheuristic optimizations

The first algorithm, inspired by the foraging behaviour of ants, was proposed in 1992 by Marco Dorigo in his doctoral dissertation. This algorithm represents various metaheuristic optimizations and belongs to the family of ant colony algorithms used in swarm intelligence techniques. Its goal was to find the shortest path in a graph. Since then, the initial idea has expanded to tackle a broader class of numerical problems, and additional challenges have evolved, each inspired by a different facet of ant behaviour. However, from a more general perspective, ACO is similar to the estimate of distribution algorithms in that it employs a model-based search strategy.

Artificial Ant

An "artificial ant" is a straightforward computational agent in ant colony optimization techniques that looks for optimal solutions to a problem needing optimization. An ant colony algorithm finds the minimum-cost path over a directed graph of weights. Each Ant initially randomly determines an order in which the graph's edges should be traversed, which is the solution for that iteration. The second part contrasts the routes taken by various ants. Finally, the pheromone concentrations at each edge are refreshed.

Pseudo Code source: javatpoint

Procedure ACO:  
    Initialize the necessary parameters and pheromone concentration;  
    while not termination do:  
        Generate initial ant population;  
        Calculate the fitness values for each Ant of the colony;  
        Find optimal solution using selection methods;  
        Update pheromone concentration;  
    end while  
end procedure  

Applications

Many derived methods have been adapted to dynamic problems in real variables, stochastic problems, multi-targets, and parallel implementations; ant colony optimization algorithms have been used to solve various combinatorial optimization problems, from quadratic assignment to protein folding and vehicle routing. Near-optimal solutions to the travelling salesman issue have also been generated using this method. When the graph is dynamic, the ant colony technique outperforms simulated annealing and genetic algorithm approaches because it may be run indefinitely and adjust to new conditions as they arise. It has applications in urban mobility and data network routing.

Conclusion

Ants reside in groups called colonies. The goal of finding food governs the ants' behaviour. Ants roam throughout their hives while searching. An ant hops from one location to another in search of food. As it moves, it leaves an organic molecule called a pheromone on the ground. Ants use pheromone trails to communicate with one another. When an ant discovers some food, it transports as much as it can. 

ACO is inspired by investigations into how ants organize themselves in the wild. The programme is based on the careful observation of real ants as they navigate to and from their nest and food sources.

Image source: Unsplash