Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!wuarchive!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!kossy.jhuapl.edu!franfh From: franfh@kossy.jhuapl.edu (Francis P. Horan) Newsgroups: comp.ai.neural-nets Subject: Re: TSP Data, me too! Keywords: TSP Message-ID: <1991Feb12.201454.26434@aplcen.apl.jhu.edu> Date: 12 Feb 91 20:14:54 GMT References: Sender: news@aplcen.apl.jhu.edu (USENET News System) Reply-To: franfh@kossy.jhuapl.edu.UUCP (Francis P. Horan) Distribution: comp Organization: The Johns Hopkins University Applied Physics Laboratory Lines: 83 In article check@sys.titech.ac.jp (Takashi Sugawara) writes: > >In article <1991Feb5.195412.3036@informatik.uni-erlangen.de> fritzke@immd2.informatik.uni-erlangen.de (B. Fritzke) writes: > > >>I'm looking for sample Traveling Salesman Problems (TSP) to test the network > >>model I have developed. > >>Specifically I look for instances of the Euclidean TSP. This is the case, > >>when all the n cities are distributed in the unit square and the distance > >>measure is the Euclidiean distance. > >> > >>I would like to get together with the examples the shortest known path > >>length or (even better) the optimal path length. The table of several TSP city arrangements at the end of this post has data in it used in the 3 references below: [1] "Computational-Complexity Reduction for Neural Network Algorityhms" by Allon Guez, Moshe Kam, James Eilbert. IEEE Transactions on Systems, Man, and Cybernetics, Vol 19, No. 2, March/April 1989. [2] " 'Neural' Computation of Decisions in Optimization Problems". Biological Cybernetics, Vol 52, pp 141-152, 1985. [3] "An effective heuristic algorithm for the travelling-salesman problem," Operation Research., vol 21, pp 48-516, 1973. All of the TSP city arrangements listed below were studied in paper [1]. The "10 City Hopfield arangement" was first described in paper [2], and the the "30 City Lin and Kernighan arangement" was first described paper [3]. We got the data points for these two arrangements over the phone from Hopfield's secretary. The columns are labelled by the top row. Each label has a number corresponding to the number of cities for the problem. Other text describes the axis (X or Y) and the city arrangement. X5 = X coordinates, Square 5 City problem Y5 = Y coordinates, Square 5 City problem X7 = X coordinates, Square 7 City problem Y7 = Y coordinates, Square 7 City problem X10Sq = X coordinates, Square 10 City problem Y10Sq = Y coordinates, Square 10 City problem X30 = X coordinates, 30 City, Lin and Kernighan arrangement, ref [3] Y30 = Y coordinates, 30 City, Lin and Kernighan arrangement, ref [3] X10Hop = X coordinates, 10 City, Hopfield and Tank arrangement, ref [2] Y10Hop = Y coordinates, 10 City, Hopfield and Tank arrangement, ref [2] X5 Y5 X7 Y7 X10Sq Y10Sq X30 Y30 X10Hop Y10Hop 0.0 1.0 0.0 1.0 0.143 1.0 .4384 .692 .22926 .76688 .5 1.0 .25 1.0 0.286 1.0 .4232 .2328 .68318 .50919 1.0 1.0 .50 1.0 0.429 1.0 .7186 .6939 .87456 .64464 1.0 0.0 .75 1.0 0.571 1.0 .3956 .1845 .84747 .35396 0.0 0.0 1.0 1.0 0.714 1.0 .9529 .4058 .39889 .45709 1.0 0.0 0.857 1.0 .6321 .0704 .23631 .13318 0.0 0.0 1.000 1.0 .6094 .2125 .16605 .22603 0.0 1.0 .3693 .5692 .66246 .25021 0.0 0.0 .3325 .6035 .61770 .26247 1.0 0.0 .0774 .8135 .51267 .93921 .5412 .1743 .3966 .1180 .2036 .4527 .7645 .8556 .5043 .0289 .9983 .0065 .6888 .8236 .6012 .6401 .5931 .6716 .7744 .7172 .3720 .8944 .2682 .4146 .6178 .2461 .6836 .5692 .6604 .5272 .6240 .5331 .4605 .8192 .353 .445 .3808 .2468 .8341 .3587