<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>05476549</id>
  <dt>j</dt>
  <an>05476549</an>
  <augroup>
    <au>Moody, Shirley A.</au>
  </augroup>
  <ti>Incoroporating procedural knowledge withing a declarative modelling system.</ti>
  <so>J. Intell. Syst. 6, No. 1, 45-62 (1996).</so>
  <py>1996</py>
  <pu>Freund Publishing House Ltd., London</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
    <cc>I.6</cc>
  </ccgroup>
  <utgroup>
    <ut>linear programming</ut>
    <ut>procedural knowledge</ut>
    <ut>algebraic modelling languages</ut>
    <ut>multi-commodity network flow</ut>
  </utgroup>
  <cigroup>
  </cigroup>
  <ligroup>
    <li>doi:10.1515/JISYS.1996.6.1.45</li>
  </ligroup>
  <abgroup>
    <ab>Summary: It is well established that Mathematical Programming (MP) problems are modelled within a declarative framework. Consequently algebraic modelling languages are widely accepted in the mathematical programming community as important tools for constructing and analyzing MP problems. There is a large class of models, however, in which the coefficients of the constraint matrix cannot be specified immediately in the declarative form using the raw problem data. In these cases many analysts abandon modelling languages and revert to writing computer programs. The paper addresses this issue by extending a modelling language to link the declarative model statement with the necessary procedures for specifying the columns of the technology matrix. This has the advantage that a considerable amount of the problem owner's domain knowledge can be captured in the modelling phase.</ab>
    <rv></rv>
  </abgroup>
</item>