documents/intro_paper/eclipse.tex

79 In \ECL the same constraint can be treated concurrently by several
142 In the next section we shall explore
150 In the following two  sections we shall examine the \ECL facilities for
152 In \cite{consprog} we
198 In this case the new choice is typically compared with the previous
229 In \ECL each decision is modelled by a variable, and each choice
275 a logical form of loops. In the example, it is equvalent to
340 In other words there are constructs which cannot be mapped onto any
394 In mathematical programming the model is essentially flat, which not
431 In this case, if also $B_{r2}=0$ then $R_1 \geq R_2 + 1$,
438 In \ECL this constraint can be modelled directly in logic, as
523 In practice packages providing an embedding of constraints in
541 In fact the problem of embedding has been overcome, though first
567 In our first example we modelled a map colouring problem using only
606 In the rest of the article, ``query {\em N}'' always refers to the query
844 In figure \pageref{fdcoins} we illustrate an \ECL program for solving the
876 In this case the \verb0#=0 and \verb0#>=0 constraints and the
937 In figures \ref{mapconceptual} and \ref{mapdesign}, above,  we showed a
950 In particular it is crucial during debugging!
1017 In response \ECL prints out the variable name and its newly assigned
1053 In this case it calls {\em indomain} first on $X$, then $Y$ and then
1063 In choosing the next decision to make, when solving a problem, it is
1068 In the example there are three variables, $X$, $Y$ and $Z$
1134 In fact, internally, finite domains are stored as lists of intervals
1158 In this example the domain of $X$ could theoretically be reduced to
1166 In general any search procedure can be substituted for {\em
1251 In query 2, the programmer enters
1270 In other words {\em ria} supports propagation of intervals, using the
1352 In the model shown in figure \pageref{housearea}, $tcost$ returns the
1410 In cases like this, propagation can be augmented by a technique known
1445 In this case the system failed to detect the
1474 In the discussion so far, we have not yet mentioned an important
1478 In the travelling salesman problem for example, the decisions of what
1528 In fact MIP packages typically not only offer optimisation as a
1567 In order to establish that there is only one possible value for $X$ we
1753 In query 2, the constraint deduces that, since the tasks cannot
1756 In query 3, since the tasks are bound to overlap, the constraint
1833 In this example the constraint handling rules use finite domain
1863 In order to use constraint handling rules, it is necessary to
1974 In figure \pageref{suspensions} the first query loads the {\em fd}
1984 suspend}.  In this case the priority is $1$, which is the highest
2009 In this case the goal is suspended until the lower bound of the finite
2055 In figure \pageref{suspcons} we implement a behaviour for our {\em ndiff}
2066 In this case it imposes the constraint that $X \verb0#>=0 Y+N$.
2127 In the preceding sections we have encountered two optimisation
2141 In \ECL the default search method is depth-first search and
2293 In query 1 the {\em fdplex} is loaded, and it automatically loads the
2334 In cases where complete search is impractical, the heuristics guiding
2401 In the case of repair, however, the tentative value comes not from a
2409 In this case the tentative value violates the new constraint.
2410 In case there is no violation, of course, the tentative values
2425 In this case the tentative values do no more than support a specific
2484 The constraint $Y \verb0##0 X$ is not affected by the update. In
2519 In case the problem additionally requires some cost to be minimised,
2544 In the algorithm used for this application, the relaxed optimum is
2682 \newblock In Proceedings CP 96 Workshop on Constraint Programming Application.
2704 \newblock In {\em Proc. 12th National Conference on Artificial Intelligence},