The DirectSearch package is a powerful Maple  tool. However, every soft has its advantages and disadvantages. In particular, the DS has problems in the case of a thin feasible set in higher dimensions. Recently a serious bug in the DS was detected by me. Solving an optimization problem, the DirectSearch produces the error communication

Warning, initial point [x1 = 1., x2 = 1., x4 = 2., y1 = 2., y2 = 3., y4 = 2.] does not satisfy the inequality constraints; trying to find a feasible initial point
Error, (in DirectSearch:-Search) cannot find feasible initial point; specify a new one
 while that initial point satisfies the constraints.

 

restart

DirectSearch:-Search(((x2-x1)^2+(y2-y1)^2)*((x4-x1)^2+(y4-y1)^2), {seq(parse(y || j) >= -(2/3)*parse(x || j)+2, j = 1 .. 4), seq(parse(y || j) >= (1/2)*parse(x || j)-3/2, j = 1 .. 4), seq(parse(y || j) <= 4, j = 1 .. 4), seq(parse(y || j) <= -3*parse(x || j)+16, j = 1 .. 4), seq(parse(y || j) <= 2*parse(x || j)+2, j = 1 .. 4), (x2-x1)*(x4-x1)+(y2-y1)*(y4-y1) = 0, (x3-x2)*(x2-x1)+(y3-y2)*(y2-y1) = 0, (x4-x1)*(x4-x3)+(y4-y1)*(y4-y3) = 0, (x4-x3)*(x3-x2)+(y4-y3)*(y3-y2) = 0}, maximize, initialpoint = [x1 = 1, x2 = 1, x3 = 2, x4 = 2, y1 = 2, y2 = 3, y3 = 3, y4 = 2])

Error, (in DirectSearch:-Search) cannot find feasible initial point; specify a new one

 

eval({seq(parse(y || j) >= -(2/3)*parse(x || j)+2, j = 1 .. 4), seq(parse(y || j) >= (1/2)*parse(x || j)-3/2, j = 1 .. 4), seq(parse(y || j) <= 4, j = 1 .. 4), seq(parse(y || j) <= -3*parse(x || j)+16, j = 1 .. 4), seq(parse(y || j) <= 2*parse(x || j)+2, j = 1 .. 4), (x2-x1)*(x4-x1)+(y2-y1)*(y4-y1) = 0, (x3-x2)*(x2-x1)+(y3-y2)*(y2-y1) = 0, (x4-x1)*(x4-x3)+(y4-y1)*(y4-y3) = 0, (x4-x3)*(x3-x2)+(y4-y3)*(y3-y2) = 0}, [x1 = 1, x2 = 1, x3 = 2, x4 = 2, y1 = 2, y2 = 3, y3 = 3, y4 = 2])

{0 = 0, -1 <= 2, -1 <= 3, 2 <= 4, 2 <= 6, 2 <= 10, 2 <= 13, 3 <= 4, 3 <= 6, 3 <= 10, 3 <= 13, -1/2 <= 2, -1/2 <= 3, 2/3 <= 2, 2/3 <= 3, 4/3 <= 2, 4/3 <= 3}

(1)

``

 

Download opti.mw


Please Wait...