On the optimum capacity of capacity expansion problems (Q2472185): Difference between revisions

From MaRDI portal
Set OpenAlex properties.
ReferenceBot (talk | contribs)
Changed an Item
 
Property / cites work
 
Property / cites work: Improving the location of minisum facilities through network modification / rank
 
Normal rank
Property / cites work
 
Property / cites work: Bottleneck Capacity Expansion Problems with General Budget Constraints / rank
 
Normal rank
Property / cites work
 
Property / cites work: Weight reduction problems with certain bottleneck objectives. / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4875228 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Approximation algorithms for certain network improvement problems / rank
 
Normal rank
Property / cites work
 
Property / cites work: A Faster Strongly Polynomial Minimum Cost Flow Algorithm / rank
 
Normal rank
Property / cites work
 
Property / cites work: On budget-constrained flow improvement. / rank
 
Normal rank
Property / cites work
 
Property / cites work: An inverse maximum capacity path problem with lower bound constraints / rank
 
Normal rank
Property / cites work
 
Property / cites work: A capacity expansion problem with budget constraint and bottleneck limitation / rank
 
Normal rank
Property / cites work
 
Property / cites work: Some reverse location problems / rank
 
Normal rank
Property / cites work
 
Property / cites work: A class of bottleneck expansion problems / rank
 
Normal rank

Latest revision as of 16:22, 27 June 2024

scientific article
Language Label Description Also known as
English
On the optimum capacity of capacity expansion problems
scientific article

    Statements

    On the optimum capacity of capacity expansion problems (English)
    0 references
    0 references
    0 references
    0 references
    20 February 2008
    0 references
    The authors consider several capacity expansion problems. Given a finite set \(E=\{e_1, \ldots, e_n\}\) with a capacity \(c_i\) for each \(e_i\) and a family \({\mathcal F}\) of subsets \(F\) of \(E\), the goal is to find an optimal capacity value \(r^*\) such that the efficiency \(f(r^*)\) minus the cost \(\Phi(r^*)\) of the expansion is maximized. They consider several cases. They show for the case of bottleneck expansion with sum cost function the problem can be solved efficiently if an efficient algorithm for the maximum weighted capacity problem is available. For the case of sum expansion with sum cost and bottleneck expansion and bottleneck cost, similar results are shown. Finally the authors investigate the capacity expansion problem for maximum flow and show that it can bes solved in polynomial time.
    0 references
    Capacity expansion
    0 references
    bottleneck
    0 references
    maximum flow
    0 references

    Identifiers