gyang274 / gbp Goto Github PK

I am using the gbp::bpp_solver. I am getting the following error :
'''error: sort_index(): detected NaN
Error in gbp::bpp_solver_dpp_wrapper(lkit, lkbn) :
sort_index(): detected NaN'''
To calrify, I have sorted oid and bins are also in format as required. Here is the order which is giving me the error.
error_order.xlsx

Platform : Microsoft Windows 10 Enterprise OS: 10.0.16299 (Build 1709)

Hi gyang274:

I'm on a project to solve a bin packing problem. and I tried you Rscript GBP. I feels the packing works great. Thanks for that.

however I encountered several problems, plz advice, thanks ahead

1. is there a way that I can defined the quantity of the orders and containers? Because in my projects I need to pack multiple produces with the same size. and multiple same sized containers
2. can i define, dont let the products rotate, cause some products had to be face up.
3. can i define packing priority, cause some sku need to go first

I know i have many questions, thanks ahead, again, this is a great package, deeply appreciated

best

Ning

If I run the bin-packing algorithm gbp::bpp:solver with an item that does not completely fit within the dimensions of the largest bin, the program freezes and the R session is stuck/unresponsive

I recently tried to install this package on my Linux machine and came across the following error:

g++ -std=gnu++11 -I/usr/share/R/include -DNDEBUG  -I"/home/sebastian/R/x86_64-pc-linux-gnu-library/3.4/Rcpp/include" -I"/home/sebastian/R/x86_64-pc-linux-gnu-library/3.4/RcppArmadillo/include"   -I../inst/include -fpic  -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -g -c RcppExports.cpp -o RcppExports.o
g++ -std=gnu++11 -I/usr/share/R/include -DNDEBUG  -I"/home/sebastian/R/x86_64-pc-linux-gnu-library/3.4/Rcpp/include" -I"/home/sebastian/R/x86_64-pc-linux-gnu-library/3.4/RcppArmadillo/include"   -I../inst/include -fpic  -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -g -c bpp.cpp -o bpp.o
In file included from bpp.cpp:1:0:
gbp_u.h: In function ‘bool approx_equal_cpp(const T&, const T&, double)’:
gbp_u.h:26:10: error: ‘approx_equal’ is not a member of ‘arma’
   return arma::approx_equal(lhs, rhs, "absdiff", tol);
          ^
/usr/lib/R/etc/Makeconf:168: recipe for target 'bpp.o' failed
make: *** [bpp.o] Error 1

I managed to fix it by updating my version of RcppArmadillo. Apparently the approx_equal function was only introduced in 0.6.700.3.0 which was released 2016-04-05.

The package should declare this requirement somewhere. Perhaps in the DESCRIPTION?

There are 28 items and the container "K0004" is able to fit all of them. However, gbp couldn't fit the last item (1000, 1000, 1000, 100), it can only fit 27/28.

The "breaking point" of the last item is (877, 1000, 1000, 100) which if the value of 1000 is changed to 877, it packs all items correctly.

bn <- data.table::data.table(
 id = c("K0001", "K0002", "K0003", "K0004", "K0005"),
 l  = c(06.0000, 10.0000, 09.0000, 1500.0000, 22.0000),
 d  = c(06.0000, 08.0000, 08.0000, 1000.0000, 14.0000),
 h  = c(06.0000, 06.0000, 07.0000, 3200.0000, 09.0000),
 w  = c(600.000, 600.000, 800.000, 12000.000, 800.000)
)

it <- data.table::data.table(
 l   = c(  75, 600, 720, 611, 611,  656,  656, 100, 700,  80,  80,  40,  40,  40,   16,   16,   18, 2500, 700,   75, 100, 720,  55, 750, 535, 100, 700, 1000),
 d   = c( 650, 150, 620,  31,  31,   18,   18, 850, 400, 770, 770, 100, 100, 100, 2500, 2500, 2720,  650, 400,  650, 650, 620, 500,  17, 110, 550, 500, 1000),
 h   = c(1600,  18,  78, 791, 799, 2033, 2033, 750,  50, 850, 850, 165, 165, 165,   11,   11,  160,   30,  50, 1600, 750,  78, 745,  30, 500, 750,  50, 1000),
 w   = c(3500,  14, 130,  99,  99,  143,  143, 250,  60, 150, 150,   1,   1,   1,    4,    4,   48,  210,  60,  350, 220, 130,  90,  26,  35, 200,  75,  100)
)

ldhw <- t(as.matrix(it))

m <- t(as.matrix(bn[ , .(l, d, h, w)])) # multple bin
p <- gbp4d_solver_dpp_prep_create_p(ldhw, m[, 4L]) # single bin
sn4d <- gbp4d_solver_dpp(p, ldhw, m[, 4L])
gbp4d_viewer(sn4d)

Thanks for the package!

Would you consider isolating the very packing algorithm in a separate package from the rgl-based viewer? Or at least make rgl as Suggested not Imported?

rgl introduces rather heavy set of dependencies. If just want to use (a package that uses) your implementation internally, and, say, on a headless server, I need to install rgl which requires a lot of system libraries including X11....

bn l=1870 d=1200 h=1000 w=1000000
12 times it: l=426 d=795 h=390 w=1.12

those 12 items trivially fit into the bn, but the solution can only pack 8 in one bn and adds a second bin. why?

Can you add pdf file or link with a detailed description of the algorithm?

Hi. Can I somehow turn off rotation and get a solution without rotation of objects? thanks. Ivo