Solver foundation simplex solution takes a lot of time

Question

Trying to solve the following problem using Solver foundation:

Given: range: {x | x from double} and points: {(x,y) | x,y from double}

Find piece wise linear function - { (a,b) | a,b from double} where:

• Linear lines drawn are between the x'es from range.
• Covers all the points.
• Minimal area under the graph.

Example: range: {1, 2, 3} , points {(1,40), (1.5,40), (2.5,70)}

My solution:

Minimize the following problem with Simplex:

foreach i range add :

var ai = new Decision(Domain.RealRange(0, 100), null);
var bi = new Decision(Domain.RealRange(0, 100), null);
model.AddDecisions(a, b);

foreach point from points that fall in i range add constraint

model.AddConstraints("c{0}".F(pointIdx), a * point.x + b >= point.y);

Then add goal:

model.AddGoal("area", GoalKind.Minimize, goal);

And get the solution:

var solution = context.Solve(new SimplexDirective());

The solution works gives me a right answer but it takes a lot of time for simple case it takes 130 ms. Could you tell me what I am doing wrong? Where can I optimize ? Is Simplex method right for this case? Do we have better software solution then SolverFoundation for optimization.

Answer 1

try loading your model from OML string, enforce simplex solver to check if your problem is LP