Virtual Ants Research Articles

A non-linear estimation and model predictive control algorithm based on ant colony optimization

A new heuristic controller, called the continuous ant colony controller, is proposed for non-linear stochastic Gaussian/non-Gaussian systems. The new controller formulates the state estimation and the model predictive control problems as a single stochastic dynamic optimization problem, and utilizes a colony of virtual ants to find and track the best estimated state and the best control signal. For this purpose, an augmented state space is defined. An integrated cost function is also defined to evaluate the points of the augmented state space, explored by the ants. This function minimizes simultaneously the state estimation error, tracking error, control effort and control smoothness. Ants search the augmented state space dynamically in a similar scheme to the optimization algorithm, known as the continuous ant colony system. Performance of the new model predictive controller is evaluated for three non-linear problems. The problems are a non-linear continuous stirred tank reactor, a non-linear cart and spring system, and the attitude control of a non-linear quadrotor. The results verify successful performance of the proposed algorithm from both estimation and control points of view.

Biologically inspired ant colony simulation

AbstractWe present a unified biologically inspired approach to simulate ant colonies inspired by the key observation of collective behaviors of ants in nature. To generate the trajectories of virtual ants, we construct a motion controller to determine the motion states and the paths of virtual ants, considering dynamic internal and external interactions. The motion controller computes a target position for each ant at every time step according to its motion states. The motion states include four states: basic movement, the stop state, and two dynamic interactions (i.e., internal and external, respectively referring to interaction with neighbors for necessary information transfer about the destination, and interaction with surroundings such as food sources, nests, and obstacles) to represent basic exploration, casual or intentional stop, and purposeful movement, respectively. Based on the motion states, the motion controller plans an optimal path for each virtual ant. Through many simulation experiments, we demonstrate that our method is controllable, scalable, and flexible to simulate hybrid colonies with a large number of ants.

A hybridization of extended Kalman filter and Ant Colony Optimization for state estimation of nonlinear systems

In this paper, a new nonlinear heuristic filter based on the hybridization of an extended Kalman filter and an ant colony estimator is proposed to estimate the states of a nonlinear system. In this filter, a group of virtual ants searches the state space stochastically and dynamically to find and track the best state estimation while the position of each ant is updated at the measurement time using the extended Kalman filter. The performance of the proposed filter is compared with well-known heuristic filters using a nonlinear benchmark problem. The statistical results show that this algorithm is able to provide promising and competitive results. Then, the new filter is tested on a nonlinear engineering problem with more than one state. The problem is to estimate simultaneously the states of an unmanned aerial vehicle as well as the wind disturbances, applied to the system. In this case, a processor-in-the-loop experiment is also performed to verify the implementation capability of the proposed approach. This paper also investigates the real-time implementation capability of the proposed filter in the attitude estimation of a three degrees of freedom experimental setup of a quadrotor to further investigate its effectiveness in practice.

Simulating collective transport of virtual ants

AbstractThis paper simulates the behaviour of collective transport where a group of ants transports an object in a cooperative fashion. Different from humans, the task coordination of collective transport, with ants, is not achieved by direct communication between group individuals, but through indirect information transmission via mechanical movements of the object. This paper proposes a stochastic probability model to model the decision‐making procedure of group individuals and trains a neural network via reinforcement learning to represent the force policy. Our method is scalable to different numbers of individuals and is adaptable to users' input, including transport trajectory, object shape, external intervention, etc. Our method can reproduce the characteristic strategies of ants, such as realign and reposition. The simulations show that with the strategy of reposition, the ants can avoid deadlock scenarios during the task of collective transport.

Application of ant colony algorithm to geochemical anomaly detection

The methods for geochemical anomaly detection are commonly based on statistical models that require assumption of the sample population to satisfy a particular distribution. In practice, the assumption of a particular distribution may degrade the performance of geochemical anomaly detection. In this paper, an ant colony algorithm is used to detect geochemical anomalies. The new method does not require assumption that the geochemical data satisfy a particular distribution. Applying this method to detect geochemical anomalies, we only need to put a number of “virtual ants” randomly into a geochemical grid map and let each ant complete its iterative search process. When the algorithm gets converged, the ants tend to aggregate in the geochemical anomalous regions where geochemical element concentration values are significantly greater than surrounding background. The number of times each grid point is visited by ants can be recorded in ant density data for geochemical anomaly identification. The ant density data are almost not affected by regional variations of geochemical background, thus they are suitable for identifying geochemical anomalies using a threshold method. As an illustration, the ant colony algorithm is applied to detect geochemical anomalies in interpolated concentration data of Au, Ag, Cu, Pb, and Zn in the Altay district in northern Xinjiang in China. The results show that the ant colony algorithm can properly identify geochemical anomalies. Anomalies detected by the ant colony algorithm occupy 9.5% of the study area and contain 36% of the known mineral deposits; and anomalies identified using the Youden index method occupy 16.4% of the study area and contain 56% of the known mineral deposits. Therefore, the ant colony algorithm can serve as a feasible swarm intelligence paradigm for geochemical anomaly detection.

ANT STRATEGIES FOR NEGOTIATING OBSTACLES

Ants are notoriously good at finding the most direct route to food. Laying down volatile pheromones to guide their colleagues, ant colonies eventually home in on the most direct path as pheromone is lost from longer paths and added to more heavily used shorter routes. Finding the optimal paths through networks is also a challenge for human delivery agents, telephone routers and systems analysts, so software engineers have turned to trail-blazing ants for inspiration to solve these problems. Chris Reid and Madeleine Beekman from Sydney University, Australia, and David Sumpter from Uppsala University, Sweden, explain that one network-solving algorithm uses virtual ants to lay down virtual pheromone trails to identify the most direct route through, but it doesn't do well when networks continually change, as they do in real life. However, ants have faced constantly changing landscapes for millions of years, so Reid and his co-workers decided to find out how Argentine ants identify the optimal route through a changing network when they presented the insects with the Towers of Hanoi puzzle (p. 50).Of course the team didn't ask the ants to start moving puzzle discs between pegs. Instead, they converted the task into a graph of possible moves and converted that in turn into a maze of hexagons. Having initially allowed a colony of ants to explore the maze and lay down pheromones (short-lived foraging pheromone and long-lasting exploration pheromone), the team then placed food at the end of the maze and filmed the ants as they tried to identify the optimal route through the maze to the food while the team blocked off routes and opened up alternatives.The ants had little problem coping with change. Initially, the insects found the shortest route, scurrying along the outer edge of the maze to the food source. However, when the team blocked this route and opened up another through the middle of the maze, the ants changed course, zigzagging across the maze, perpendicular to their original route until they found the alternative. The team suspects that both pheromones played critical roles, with short-lived foraging pheromone allowing the ants to respond quickly when the route changed and long-lived exploration pheromone rapidly stabilising the optimal route. Also, when an ant encountered an obstacle, Reid and his colleagues noticed that she began searching for a route around the blockage, rather than retracing her steps. The team suspects that the ants don't just rely on geometry to identify paths and negotiate obstacles; they use an internal compass to help them reach their goal. Finally, the trio hopes that their new discovery that ants use multiple pheromones and possibly compasses will inspire software engineers to develop alternative ant-based algorithms to solve complex dynamic network problems.

Artificial pheromone for path selection by a foraging swarm of robots

Foraging robots involved in a search and retrieval task may create paths to navigate faster in their environment. In this context, a swarm of robots that has found several resources and created different paths may benefit strongly from path selection. Path selection enhances the foraging behavior by allowing the swarm to focus on the most profitable resource with the possibility for unused robots to stop participating in the path maintenance and to switch to another task. In order to achieve path selection, we implement virtual ants that lay artificial pheromone inside a network of robots. Virtual ants are local messages transmitted by robots; they travel along chains of robots and deposit artificial pheromone on the robots that are literally forming the chain and indicating the path. The concentration of artificial pheromone on the robots allows them to decide whether they are part of a selected path. We parameterize the mechanism with a mathematical model and provide an experimental validation using a swarm of 20 real robots. We show that our mechanism favors the selection of the closest resource is able to select a new path if a selected resource becomes unavailable and selects a newly detected and better resource when possible. As robots use very simple messages and behaviors, the system would be particularly well suited for swarms of microrobots with minimal abilities.

The MAGIC-5 CAD for nodule detection in low dose and thin slice lung CTs

Lung cancer is the leading cause of cancer-related mortality in developed countries. Only 10–15% of all men and women diagnosed with lung cancer live 5 years after the diagnosis. However, the 5-year survival rate for patients diagnosed in the early asymptomatic stage of the disease can reach 70%. Early-stage lung cancers can be diagnosed by detecting non-calcified small pulmonary nodules with computed tomography (CT). Computer-aided detection (CAD) could support radiologists in the analysis of the large amount of noisy images generated in screening programs, where low-dose and thin-slice settings are used. The MAGIC-5 project, funded by the Istituto Nazionale di Fisica Nucleare (INFN, Italy) and Ministero dell’Università e della Ricerca (MUR, Italy), developed a multi-method approach based on three CAD algorithms to be used in parallel with a merging of their results: the Channeler Ant Model (CAM), based on Virtual Ant Colonies, the Dot-Enhancement/Pleura Surface Normals/VBNA (DE-PSN-VBNA), and the Region Growing Volume Plateau (RGVP). Preliminary results show quite good performances, to be improved with the refining of the single algorithm and the added value of the results merging.

Langton's flies

Langton's ant is an automaton defined over a two-dimensional grid. Its movement is governed by the environment in a very simple way: it turns to the left over white vertices and to the right over black vertices. This definition only applies on a two-dimensional space. We look for generalizations of this automaton to n-dimensional lattices. Remembering the different ways through which the ant was originally defined, we consider two approaches: the first comes from physics (lattice gas) and the second from artificial life (virtual ants). Two generalizations are proposed defining two families of dynamical systems. From the physics point of view, the ant is seen as a particle and hence it has no internal state other than its velocity. From the artificial life point of view, the ant is viewed as an insect, and it has an orientation in space which we represent by an orthogonal basis. This constitutes the ant's internal state. This formulation allows us to define the ant's behaviour without drawing upon any information relative to the global system of external coordinates. Each model yields different sets of rules with distinctive behaviours. We characterize all the possible rules satisfying some basic restrictions. We found that many rules produce trajectories which are restricted to a diagonal plane and are equivalent to a version of Langton's ant over a two-dimensional grid, squared or hexagonal. In the particle model, only two of them use the whole space, and it is shown that such rules do not admit periodical trajectories. This result reinforces a previous one reported by Leonid Bunimovich, who states that ‘the skeleton of any bounded trajectory cannot contain any three-dimensional polyhedron’.

Evolving strategy for a probabilistic game of imperfect information using genetic programming

We provide the complete record of methodology that let us evolve BrilliAnt, the winner of the Ant Wars contest. Ant Wars contestants are virtual ants collecting food on a grid board in the presence of a competing ant. BrilliAnt has been evolved through a competitive one-population coevolution using genetic programming and fitnessless selection. In this paper, we detail the evolutionary setup that lead to BrilliAnt's emergence, assess its direct and indirect human-competitiveness, and describe the behavioral patterns observed in its strategy.

Designing conducting polymers using bioinspired ant algorithms

Ant algorithms are inspired in real ants and the main idea is to create virtual ants that travel into the space of possible solutions depositing virtual pheromone proportional to how good a specific solution is. This creates an autocatalytic (positive feedback) process that can be used to generate automatic solutions to very difficult problems. In the present work we show that these algorithms can be used coupled to tight-binding Hamiltonians to design conducting polymers with pre-specified properties. The methodology is completely general and can be used for a large number of optimizations problems in materials science.

Optimal product design using a colony of virtual ants

The optimal product design problem, where the “best” mix of product features are formulated into an ideal offering, is formulated using ant colony optimization (ACO). Here, algorithms based on the behavior of social insects are applied to a consumer decision model designed to guide new product decisions and to allow planning and evaluation of product offering scenarios. ACO heuristics are efficient at searching through a vast decision space and are extremely flexible when model inputs continuously change. When compared to complete enumeration of all possible solutions, ACO is found to generate near-optimal results for this problem. Prior research has focused primarily on optimal product planning using consumer preference data from a single point in time. Extant literature suggests these formulations are overly simplistic, as a consumer’s level of preference for a product is affected by past experience and prior choices. This application models consumer preferences as evolutionary, shifting over time.

An ant colony optimization approach to addressing a JIT sequencing problem with multiple objectives

This research presents an application of the relatively new approach of ant colony optimization (ACO) to address a production-sequencing problem when two objectives are present — simulating the artificial intelligence agents of virtual ants to obtain desirable solutions to a manufacturing logistics problem. The two objectives are minimization of setups and optimization of stability of material usage rates. This type of problem is NP-hard, and therefore, attainment of IP/LP solutions, or solutions via complete enumeration is not a practical option. Because of such challenges, an approach is used here to obtain desirable solutions to this problem with a minimal computational effort. The solutions obtained via the ACO approach are compared against solutions obtained via other search heuristics, such as simulated annealing, tabu search, genetic algorithms and neural network approaches. Experimental results show that the ACO approach is competitive with these other approaches in terms of performance and CPU requirements.