Question

所以我已经解决了这个问题一段时间了...... 我想用clpfd库解决Prolog中的这个优化问题：想象一下，公司有几个他们想要实施的措施，每个措施都会对某些标准产生影响并定义成本。该公司有一个限制buget。例如，一系列措施:(每项措施的5个标准） [[id，cost，[每个标准的影响列表（1,2,3,4,5）从0到100]，平均影响]，......]

measure(1, 100, [50,50,50,60,50]).

这是一个例子，我创建一个具有平均影响的度量列表列表......

我需要的是一种方法，它将检查我是否可以加入总预算低于预算的措施，并寻找最佳措施来改善对所有标准的影响。

以下是小型数据库的代码和示例：

measure(1, 100, [100,90,100,100,100]).
measure(2, 10, [50,70,80,80,80]).
measure(3, 200, [50,50,50,100,20]).
measure(4, 80, [100, 100,100,100,100]).

输入将是：

strategy([0.2,0.2,0.2,0.2,0.2],190).

我想要的解决方案是：措施[4]。因为平均影响较高且成本低于预算。但我继续得到第一个解决方案

strategy(Priorities, Budget):-
sumlist(Priorities, 1.0),
length(Priorities, 5),

setof([Id, Cost, Criteria], measure(Id, Cost, Criteria), List),
length(List, MaxLength),

%List of usable measures (Cost < Budget)
setof([Id, Cost, Criteria, Impact], (measure(Id, Cost, Criteria), Cost < Budget, sum_prod(Criteria, Priorities, Impact)), MeasuresList),
TotalImpact in 0 .. 100,

Length in 1 .. MaxLength,
length(Measures, Length),
domain(Measures, 1, MaxLength),
all_different(Measures),
write('Length: '), write(Length), nl,

checkMeasures(Measures, MeasuresList, Budget, TotalImpact),
labeling(maximize(TotalImpact), Measures),
write('TotalImpact: '), write(TotalImpact), nl,
write('Measures: '), write(Measures).

checkMeasures(Measures, MeasuresList, Budget, TotalImpact) :-
    measuresClause(Measures, MeasuresList, 0, Budget, 0, TotalImpact).

measuresClause([Id|Rest], MeasuresList, TotalSpent,
                Budget, ImpactSum, TotalImpact):-
    ImpactSum == 0,
    member([Id, Cost, _, Impact], MeasuresList),
    I is round(Impact),
    Aux #= I + ImpactSum,
    Sum #= Aux,
    TotalBudget #= Cost + TotalSpent,
    measuresClause(Rest, MeasuresList, TotalBudget, Budget, Sum, TotalImpact).
measuresClause([Id|Rest], MeasuresList, TotalSpent,
                Budget, ImpactSum, TotalImpact) :-
    member([Id, Cost, _, Impact], MeasuresList),
    I is round(Impact),
    Aux #= I + ImpactSum,
    Sum #= Aux/2,
    TotalBudget #= Cost + TotalSpent,
    measuresClause(Rest, MeasuresList, TotalBudget, Budget, Sum, TotalImpact).
measuresClause([], _, TotalSpent, Budget, ImpactSum, TotalImpact) :-
    TotalSpent #=< Budget,
    TotalImpact #= ImpactSum.

我试过这样做：

    labeling([max(TotalImpact),time_out(30000,_), all], Measures),

但我刚收到第一个结果

约束序言，创建优化方法

0 个答案: