In St. Petersburg, researchers from the ITMO University undertook the development of the theoretical foundations of dynamic parameters in evolutionary algorithms. This breakthrough direction in programming and information technology, which will create an entirely new logistics model, will improve the quality of machine learning.
the Term “evolutionary algorithms” – optimization using algorithms. The creators of this idea were inspired by Darwin’s theory. In the animal world survival of the adapted, and in the world of algorithms most often used choices that lead to optimum results. Consequently, algorithms that are ineffective, over time, cease to be used and breakthrough solutions to improve and be even stronger and improve processes. It turns out the same kind of evolution.
As they say at the ITMO University, the efficiency of algorithms depends to a large extent from their settings and original settings at startup. To use only mathematical methods are impossible, too much data is used, and the tasks overly complex. Poverty theoretical framework significantly reduces the use of evolutionary algorithms.
Accordingly, the St. Petersburg researchers have set ourselves the task to obtain a rigorous mathematical proof of efficiency of evolutionary algorithms at different settings. The scientists also intend to develop new methods settings, to conduct computational experiments and present the results in the form of program code.
In particular, scientists will be able to perform tasks with dynamic properties. Met with them at the time, each. For example, travel time from work to home can vary considerably depending on traffic conditions. In logistics companies, where the route consists of several points, the lack of attention to road conditions may result in loss of time and as a result, great losses. But such a dynamic problem as the road situation can also be fit into the evolutionary algorithm. Of course, for each individual route calculations, however, in the course of the research experts plan to make recommendations for use in such conditions.