Universal procedure for constructing a Pareto set

DOI10.1134/S0965542517010122MaRDI QIDQ2357117zbMATH OpenOpenAlexFDO

Authors Ya. I. Rabinovich

Publication date 19 June 2017

Published in Computational Mathematics and Mathematical Physics (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1134/s0965542517010122

zbMATH Keywords

(weak) effective solutions\(\omega\) concavityfinite dimensionalGermeier convolutionmultiobjective optimizationnonnegative orthantparametric construction of Pareto setuniversal procedure

Mathematics Subject Classification ID

Convexity of real functions of several variables, generalizations (26B25) Multiobjective variational problems, Pareto optimality, applications to economics, etc. (58E17) Numerical mathematical programming methods (65K05) Multi-objective and goal programming (90C29) Sensitivity, stability, parametric optimization (90C31)

Review text

Considering a multiobjective problem in finite-dimensional spaces with a continuous objective \(w:\mathbb R^s \rightarrow \mathbb R^m\) over a compact feasible set \(X\subset \mathbb R^s\) a general procedure for determining the (weak) Pareto set is theoretically proposed. Points of the Pareto set are approximated (universal procedure in Section 3) by using the scalar Germeier convolution \(f_\lambda\) (see below) with parameters \(\lambda\) in the standard simplex \(\Lambda\) and some fictive interior point method for the related scalar problems. A convergence theorem is given under Lipschitz continuity and coordinate wise \(\omega\) concavity (see below) of \(w\) and convexity of \(X\). There are no examples, no numerical computations and no complexity considerations. \(f:A\rightarrow \mathbb R\), \(A\) convex, is called \(\omega\)-concave with some \(\omega\in [0,1]\) iff \(f(y)\geq f(x)\) for all \(x,y\in A\) and convex combinations \(z\) of \(x,y\) implies \(\|y-x\| f(z)\geq \|y-z\|f(x)+\|z-x\|(\omega f(y)+ (1-\omega) f(x))\). Convexity (quasi convexity) is given for \(\omega = 1 (\omega =0)\). The Germeier convolution \(f_\lambda\) is given by \(f_\lambda (x):=\min_{1\leq k\leq m, \lambda_k>0 } \lambda_k^{-1} w_k(x)\), \(\lambda \in \Lambda\).

Recommendations

• Set contraction algorithm for computing Pareto set in nonconvex nonsmooth multiobjective optimization
• Methods of approximating Pareto set faces in a linear multiple-criteria problem
• Uniform convergence and Pareto optimality
• scientific article; zbMATH DE number 3910156
• scientific article; zbMATH DE number 7434530

Cites work

• Approximate reduction of the problem of determining a maximin to the problem of determining a maximum by means of penalty functions
• Constructing a set of effective vectors by \(\varepsilon\)-perturbations method
• scientific article; zbMATH DE number 3850843 (Why is no real title available?)
• scientific article; zbMATH DE number 3972626 (Why is no real title available?)

Cited in

(6)

• Synthesis of numerical methods for Pareto set approximation based on a universal procedure
• Construction of Pareto set in a system of conceptual uncertainty problems
• The fixed point property in convex multi-objective optimization problem
• scientific article; zbMATH DE number 4012302 (Why is no real title available?)
• scientific article; zbMATH DE number 7434530 (Why is no real title available?)
• Set contraction algorithm for computing Pareto set in nonconvex nonsmooth multiobjective optimization

This page was built for publication: Universal procedure for constructing a Pareto set

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2357117)