Solving Mixed-Integer Programming Problems Using Piecewise Linearization Methods
نویسنده
چکیده
We present an overview on piecewise linearization methods for MINLPs. This will include the concept of disjunctive constraints, which is necessary to define logarithmic reformulations of the so called disaggregated convex combination method and the convex combination method. For the case of a general univariate quadratic function we also calculate the linearization error and proof that equidistant grid points minimize this error. For a bivariate product of two variables we do the same error analysis for the case of J1-triangulations and again equidistant grid points will be optimal. The presented methods will then be applied to a newly developed model for a hybrid energy supply network problem. We show that linearizations are able to solve this non-convex optimization problem within reasonable time. Zusammenfassung Wir geben einen Überblick zu stückweisen Linearisierungsmethoden für MINLPs. Insbesondere wird dabei das Konzept der Disjunctive Constraints vorgestellt, welche notwendig sind um logarithmische Umformulierungen der sogenannten Disaggregated Concex Combination Methode und der Convex Combination Methode durchzuführen. Für den Fall einer univariaten quadratischen Funktion berechnen wir außerdem den Linearisierungsfehler und beweisen, dass eine äquidistante Wahl der Gitterpunkte der Triangulierung diesen minimiert. Für das bivariate Produkt zweier Variablen werden wir dieselbe Fehleranalyse für den Fall, dass eine J1-Triangulierung gegeben ist, durchführen und erneut zeigen, dass äquidistante Gitterpunkte optimal sind. Danach werden die eingeführten Methoden auf das neu entwickelte Modell eines hybriden Energieversorgungsnetzwerks angewendet. Wir werden zeigen, dass dieses Problem mit Hilfe von stückweisen Linearisierungen in kurzer Zeit gelöst werden kann. 2 Solving Mixed-Integer Programming Problems Using Piecewise Linearization Methods
منابع مشابه
Coordinated Planning and Scheduling of Electricity and Natural Gas Systems
In this paper we propose a model for coordinated planning of power and natural gas systems, as a part of electricity supply chain. This model takes into account costs and constraints of both systems, and with hiring simplifications and linearization methods transforms initially nonlinear formulation to a mixed integer linear programming (MILP) problem. Natural gas would be assumed steady state,...
متن کاملSolving Single Machine Sequencing to Minimize Maximum Lateness Problem Using Mixed Integer Programming
Despite existing various integer programming for sequencing problems, there is not enoughinformation about practical values of the models. This paper considers the problem of minimizing maximumlateness with release dates and presents four different mixed integer programming (MIP) models to solve thisproblem. These models have been formulated for the classical single machine problem, namely sequ...
متن کاملComparing Two-Echelon and Single-Echelon Multi-Objective Capacitated Vehicle Routing Problems
This paper aims to compare a two-echelon and a single-echelon distribution system. A mathematical model for the Single-Echelon Capacitated Vehicle Routing Problem (SE-CVRP) is proposed. This SE-CVRP is the counterpart of Two-Echelon Capacitated Vehicle Routing Problem (2E-CVRP) introduced in the authors’ previous work. The proposed mathematical model is Mixed-Integer Non-Linear Programming (MIN...
متن کاملGlobal optimization of mixed-integer polynomial programming problems: A new method based on Grobner Bases theory
Mixed-integer polynomial programming (MIPP) problems are one class of mixed-integer nonlinear programming (MINLP) problems where objective function and constraints are restricted to the polynomial functions. Although the MINLP problem is NP-hard, in special cases such as MIPP problems, an efficient algorithm can be extended to solve it. In this research, we propose an algorit...
متن کاملA global optimization method for nonconvex separable programming problems
Conventional methods of solving nonconvex separable programming (NSP) problems by mixed integer programming methods requires adding numerous 0±1 variables. In this work, we present a new method of deriving the global optimum of a NSP program using less number of 0±1 variables. A separable function is initially expressed by a piecewise linear function with summation of absolute terms. Linearizin...
متن کامل