Using Feasibility Cuts to Accelerate the Convergence of Modes

Abstract

System for Closure Optimization Planning and Evaluation (SCOPE) is a system for closure planning developed by the Production and Distribution Research Center (PDRC) at Georgia Tech. Two types of decisions are involved in closure planning: 1) allocate the available capacity of different asset types to transportation channel, and 2) load and schedule the various movement requirements to these channels in accordance with the allocated capacities. The specific linear program associated with MRMATE is of the special type known as the minimal cost generalized network flow problem. Its special structure allows fast solution even for very large problems. However, the channel configuration created by LIFTCAP is not guaranteed to allow delivery of all the movement requirements until the final iteration. In many cases finding a feasible solution, i.e. channel configuration which allows 100% delivery, when one exists, is as hard as finding the optimal solution. The network of MRMATE is examined and special classes of cutsets (partitions of the node set of the network) is identified in this network which are themselves associated with necessary conditions for the network to admit feasible flow. Some of these conditions are then added to LIFTCAP as additional constraints in order to make the channel configuration produced by this sub) problem closer to feasibility. These cuts are thus termed 'feasibility' cuts.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1988
Accession Number
ADA198715

Entities

People

  • H. D. Ratliff
  • John J. Jarvis
  • Moshe Eben-chaime

Organizations

  • Georgia Tech

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Algorithms
  • Convergence
  • Data Sets
  • Engineering
  • Equations
  • Evolutionary Algorithms
  • Heuristic Methods
  • Industrial Engineering
  • Iterations
  • Linear Programming
  • Optimization
  • Shipping
  • Simplex Method
  • Systems Engineering
  • Transportation
  • Travel Time

Readers

  • Graph Algorithms and Convex Optimization.
  • Operations Research
  • Systems Analysis and Design