gbp2q {gbp}

gbp2q

Description

generalized bin packing problem in 2 dimension, a.k.a rectangle fill.

Usage

gbp2q

Format

An object of class C++Class of length 1.

Details

gbp2d init a profit vector p, a length vector l, a depth vector d, a length constraint ml, and a depth constraint md on l x d rectangle with geometry intepretation.

gbp2d solver would solve

maximize sum_j=1^n p_j k_j

subject to fit (l_j, d_j) at coordinate (x_j, y_j) such that no overlap in ml x md, j = 1, ...., n

and instantiate a gbp2d object with a x-axis coordinate vector x, a y-axis coordinate vector y, a selection vector k, and an objective o.

gbp2q solver would also select the most preferred often smallest m from a list of m(l, d) after determine all or the higest volume set of ld can fit into one m(l, d).

a gbp2q class instance has 7 fields:

- p: profit of it fit into bn <vector>

created via cluster max(l, d) and min(l, d) via gbp2d_solver_dpp_prep_create_p()

- it: it position and scale <matrix>

- x, y it position in the bin <numeric>

- l, d it scale along x and y <numeric>

- bn: bn scale <matrix>

- l, d bn scale along x and y <numeric>

matrix of 2 rows and each column is a single bn

should make sure bn list are sorted via volume so that the first col is the most prefered smallest bn, and also the last col is the least prefered largest and often dominant bn

should make sure no X in front of Y if bnX dominant bnY, bnX dominant bnY if all(X(l, d) > Y(l, d)) and should always prefer Y.

should make sure bn such that l >= d or vice versa.

- k: selection indicator 0, 1 on it <vector>

- f: selection indicator 0, 1, 2, 3 on bn <vector>

f in result should have no 0 and only one of 1

- o: objective achivement volumn fit in over volumn overall <numeric>

- ok: a quick indicator of all it fit into bn? <bool>

See Also

Other gbp2q: gbp2d_solver_dpp_filt, gbp2q_checkr