Читать книгу Handbook on Intelligent Healthcare Analytics - Группа авторов - Страница 31

In research papers and books on this topic, there is a wide range of GRST approaches. A lot of them are still under study and are still not sophisticated in making them attractive for developers or engineers of commercial devices that create an internal MDO structure. However, there are at least a few approaches in the “available” lists of methods utilized in publicity programs that are worth mentioning.

One is the principle of “simulated ringing” that allows atoms to form large crystals at an energy stage, at a minimum, given the differences present on the physical search pathway. The other approach is the ringing for steel and other metals. The search algorithm involves a way of randomly extracting inputs from neighboring designs and merge them according to a given set of laws as applied to the optimization issue. The key task of the virtual anneal is to have a small but not nil chance of flipping from an improved to a lower configuration. This makes it possible to break the local minimum trap at the cost of temporary design inferiority and, in the long run, pays off by going onto a new quest route that can optimize the probability of achieving an optimum overall.

The “particle swarm optimization” envisages designs in production rooms as a swarm of entities (the swarm of bees was inspired). In line with simple mathematical formulae, the swarm is moved into the design space to draw the position and velocity of all particles in the swarm to incorporate local and global information.