Authors
Semenov G.
|
Degree
|
PhD in Technique, Associate Professor, Moscow Aviation Institute (National Research University) |
|---|---|
|
E-mail
|
grigory_semenov@mail.ru |
|
Location
|
Moscow |
|
Articles
|
Mathematical model of genetic algorithm in implementation for scheduling tasks of complex technical objects during pre-production stagesThe article describes the problems of scheduling systems development. The authors demonstrate complexity
of feasible schedules generation for comprehensive technological databases. The study concluded that
the task belongs to the class of non-polynomial complexity. The article formalizes the task to plan as a single
mathematical model. Particular emphasis was placed on the fact that the mathematical model should have
the property of practical problem solving. The main task is to minimize the project development time, and
the main goal is to uniform loading of each resource in a complex system. This article proves that exact solution
of mathematical optimization is not effective since it demands huge amount of computational resources.
Authors demonstrate that genetic algorithm implementation allows to find approximate feasible solution. The
authors concluded that it is acceptable to use hybrid algorithms to solve the problem. It is proposed to use
heuristic methods to further improve the resulted mathematical model. Article also demonstrates how these
modules are integrated into technological data visualization systems developed by them. The PrPlan software
package allows interacting with existing systems in the enterprise and includes process data uploading interfaces.
It can group process operations of the same type and ensure generation of technological solutions based
on prototypes. The developed mathematical model and software will significantly optimize the technological
processes in production and ensure a high level of stability in the performance of each operation. It enforces
a significant economic growth and increase the efficiency of the enterprise in general. The developed mathematical
model was successfully tested in practice in the developed PrPlan software package. Additional optimizations
that can be implemented in the model can be equally well made in the algorithm of the software
system. This will allow us to make changes quickly, without disturbing existing processes.
Read more...
|
