Cultural Algorithms. Robert G. Reynolds. Читать онлайн. Newlib. NEWLIB.NET

Автор: Robert G. Reynolds
Издательство: John Wiley & Sons Limited
Серия:
Жанр произведения: Программы
Год издания: 0
isbn: 9781119403104
Скачать книгу
initialization step of each run occurs when the system takes the data used to establish the parameters of the simulation, and generates the given number of cones. While the number of cones is set at 150, a number of these cones are not readily visible, as they are absorbed by the larger cones. Whenever two or more cones overlap, the system is designed to take the maximum of those cones at the given points where they overlap. Those cones subsumed by larger cones in the static landscape will remain hidden throughout the entirety of the run, while those hidden cones in the dynamic landscape have a chance of becoming visible when the dynamic landscape is updated and the dimensions of the cones are altered.

      The A‐values fed into the logistics function are used to determine the relative dimensions of each cone. For a given acceptable range of values for the cones to have, each cone will be individually defined based on repeated initial calls of the logistics function as a means of seeding the function, followed by subsequent calls to the function for each newly generated cone. This means that a low A value, which results in the linear results seen in The Cones World section, used for initialization will result in a number of cones that are not terribly dissimilar from one another, the changes in their dimensions being gradual and slight. Using a higher A value, such as 3.5 which was used in these two runs, will result in subsequent calls to the logistics function returning a more chaotic frequency. For this reason, subsequently generated cones can differ dramatically from one another.

Image described by caption and surrounding text. Image described by caption and surrounding text. Image described by caption.

      Meanwhile in the dynamic landscape, it is possible to see in Figure 2.10. that they are just beginning to cluster as the agents in the static landscape did, when the first dynamic change occurs to the landscape. The agent cluster, which had previously begun congregating on the overall maximum for the dynamic landscape, is suddenly clustered near a new overall maximum, although the centroid of the cluster is not on it. But because the loosely clustered group was near the overall maximum, they were able to then cluster on the maximum and send out exploratory agents to investigate the newly changed landscape. As the A‐value for the dynamic landscape is 3.5, this means that each dynamic shift will be a radical update which can result in the total possible maximum being significantly lower as no cone approaches a height similar to those of a pre‐updated landscape. This will be displayed later in the scoring results of the agents working in the static and dynamic landscapes.

      While continuing the simulation, it is possible to view not only the agents and their network shared network topology, but also the area which each influencing knowledge source encompasses. By identifying each agent with the knowledge source which is influencing it, and then compiling the coordinates of each agent, it is possible to draw a bounding box which contains all agents that adhere to a given knowledge source. The structures of these boxes and their subsequent expansions and contractions can serve to highlight the nature of each knowledge source.

      Boxes that contract over successive steps indicate an exploitative knowledge source, such as the Situational knowledge source. These knowledge sources will tend to focus on a known best example and explore in its immediate vicinity for any possible improvement. Boxes that expand over successive steps or trend toward more encompassing sizes typically represent the explorative knowledge sources, such as the Topographical knowledge source. These knowledge sources tend to send out agents to possibly high‐scoring predicted spots based on calculations made with known data.

      While explorative search suffers from covering massive amounts of ground with limited agents, exploitative suffers from a blindness brought on from agents that only focus on the immediate. With agents sharing information between knowledge sources, however, both types of knowledge source will benefit.