Exact algorithms for solving the constrained convex combinatorial assignment problems with bounds and maximum penalty

Abstract

In this paper, we address the constrained convex combinatorial assignment problem with bounds and maximum penalty (the CCCA-BMP problem, for short). Concretely, given a set M={g1,g2,…,gm}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$M=\{g_{1},g_{2},\ldots ,g_m\}$$\end{document} of m machines (edge servers) with two bound functions l,u:M→Z0+\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$l,u: M\rightarrow Z^{+}_{0}$$\end{document}, satisfying the bound requirement l(gi)≤u(gi)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$l(g_{i})\le u(g_{i})$$\end{document} for each machine gi∈M\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$g_{i}\in M$$\end{document}, and a set J={a1,a2,…,an}\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$J=\{a_1,a_2,\ldots ,a_n\}$$\end{document} of n independent jobs (computation tasks) with a penalty function p:J→R0+\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p: J\rightarrow R^{+}_{0}$$\end{document}, each machine gi\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$g_i$$\end{document} in M receives at least li\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$l_i$$\end{document} and at most ui\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$u_i$$\end{document} jobs to execute, and each job aj\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$a_{j}$$\end{document} in J must be either executed on some machine gi∈M\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$g_i\in M$$\end{document} with its processing time cij>0\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$c_{ij}>0$$\end{document}, or rejected (executed by the cloud computing center) with its penalty pj\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p_{j}$$\end{document} that we must pay for. We consider the CCCA-BMP problem and its variation, i.e., (1) The CCCA-BMP problem is to find an assignment scheme of these n jobs that satisfies the constraints as mentioned-above, the objective is to minimize the convex combination of the total processing time of the executed jobs and the maximum penalty of rejected jobs; (2) The penalized convex combinatorial assignment problem with bounds (the PCCA-B problem, for short) is to find an assignment scheme of these n jobs that satisfies the constraints as mentioned-above, the objective is to minimize the convex combination of the total processing time of executed jobs and maximum penalty of executed jobs. As our main contributions, we present two exact combinatorial algorithms to solve the CCCA-BMP problem and the PCCA-B problem, respectively. In addition, we give numerical examples to illustrate the execution processes of the two algorithms we proposed.