On 3/15/06, Brian Schott <[EMAIL PROTECTED]> wrote: > June, > > Have you considered a classical Assignment Problem > formulation http://www.nist.gov/dads/HTML/assignment.html ? > It may take some artful crafting to fit your problem into > the AP, but it would be a start. >
In my case, the rules are somewhat loose. It may or may not fit. Anyway, I will give it a shot. > If you want to pursue that further, I may have > worked out an algorithm for it based on Keith Smillie's > help. I think the algorithm is called the Russian algorithm. Thank you. I would appreciate any guidance you may afford. > > (B=) > > On Sat, 11 Mar 2006, Mike Day wrote: > > + So which is worse? not changing tables (groups) or staying with > + more than one old neighbour. Group 0 has 5 members, so if they > + must all change group, there must be at least three in one new > + group; alternatively, if one can stay in group 0, then the no > + more than one previous neighbour rule can be met. > + > + Mike > + > + June Kim wrote: > + > + >There are 17 people. They each belonged to four groups previously. > + > > + >Group 0 had five people: > + >0 > + >1 > + >2 > + >3 > + >4 > + > > + >Group1 had four people: > + >5 > + >6 > + >7 > + >8 > + > > + >Group2, ditto: > + >9 > + >10 > + >11 > + >12 > + > > + >Group3, ditto: > + >13 > + >14 > + >15 > + >16 > + > > + >I wanted to shuffle them into new groups. But there were some constraints. > + > > + >If possible, every person should move into a new group. For example, > + >person 16 should not stay in the group 3. (The groups have physical > + >locations, uh tables, and I want to change people's location) > + > > + >It is preferable that the people meet new people at their new groups. > + >Meeting just one person in the previous grouping wouldn't be too bad. > + > > + >Person number 0, 5, 9, 10, 13, 14, 16 should be well distributed among > + >the groups. > + > > + >It is preferable that person number 1, 6, 10, 16 go into each team; > + >that is, if possible, each of them should be in different groups and > + >well distributed among the four groups. > + > > + >This was a real problem I encountered a few days ago and I solved it > + >with pencil and paper. I'd like to solve it using J but can't think of > + >a good way of solving this kind of constraints problem in J. > + > > + >Any suggestions? > + > > + >June > + >---------------------------------------------------------------------- > + >For information about J forums see http://www.jsoftware.com/forums.htm > + > > + > > + > > + > > + ---------------------------------------------------------------------- > + For information about J forums see http://www.jsoftware.com/forums.htm > + > > (B=) <----------my "sig" > > Brian Schott > Atlanta, GA, USA > schott DOT bee are eye eh en AT gee em ae eye el DOT com > http://schott.selfip.net/~brian/ > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
