TECHNOLOGY FOR CONTROLLING A FAMILY OF MULTILINE PROGRAMMING PROBLEMS
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
This paper considers a class of models in which constraint equations are described by multilinear functions. This class of models is described using a tabular form for representation of a multilinear model, previously introduced by the authors. In this case, the entire set of model constraints can be placed within a single matrix structure, which is processed in an on-screen editor. This description helps to solve a number of problems related to describing the set of admissible coordinates of a multilinear model. The concept of the universal description of a multilinear model is introduced. It allows a non-programmer user to perform structural rearrangements of multilinear models using the on-screen editor quickly. The developed description solves the problem of embedding the set of admissible coordinate values of a multilinear model into a linear polyhedron. It allows one to obtain upper bounds for criteria in multilinear optimization problems. An inverse problem of embedding a linear polyhedron into the set of admissible coordinates of a multilinear model is also considered. The complexity of creating such a construction is due to the fact that the constraints of multilinear models are defined by nonconvex nonlinear functions. A sufficient condition for the existence of an immersion of a linear polyhedron in the feasible set of values of a multilinear problem is formulated. If such an immersion construction exists, linear models can be used to obtain the feasible solutions to the multilinear problems. Construction of a feasible solution is often a complex problem in its own right in multilinear models. This is related to an important area of improper mathematical programming problems. If the linear polyhedron can be successfully immersed in the acceptable set of a multilinear problem, then linear programming algorithms can be used to solve this problem. The latter have undeniable advantages over multilinear programming algorithms. The obtained results are illustrated with examples.
Fields of Study & Subjects
Advanced Optimization Algorithms Research
Optimization and Mathematical Programming
Complexity and Algorithms in Graphs
Publication details
- Journal / source
- Informatization and communication
- Volume
- 1
- Publication date
- 1 January 2026
- Language
- EN