I am struggling with a variation of the traveling sales man problem but with a limit that the vehicle can only visit 6 or 7 customers before having to come back to depot and going back out.

backout.

Distance Matrix for Webvan Deliveries Cust Cust Cust Cust Cust Cust Cust Cust Cust Cust Cust Cust Cust Route To Travel Distance 9 10 11 1213 DC1 1000 12 DC Cust 1 Cust 2 Cust 3 Cust 4 Cust 5 Cust 6 Cust 7 Cust 8 Cust 9 Cust 10 Cust 11 Cust 12 Cust 13 8 17520 17 6 1621 115 9 17 23 227182328 22 27 912 1481420 4 14 202019 22 22 26 24 30 410001 16161620 922 228 91114 22 7| 16| 23 6 12 20 12 1000 HW3 9100010 8 1000014 9 1513 17 15 17 20 23 15 20 16 17 22 132016 914 10006 6 10004 14 20 The vehicle can only deliver to 6 or 7 customers before it must return to the depot, then it must go out and deliver to the remaining 7 or 6 4 1000 016 8 17 9 19 1611 14 0 1000 81 5 1 6 100014 19 8 14 1000 6 1812 220 1720 16 16 23 14 2219 21 28 18 26221 1122 14 24 21 15 27 200 282223 20 12 916 100013 20 6 1319 14 16 12 9131000 91620 8 1000 array is the distance matrix, 14x14 label for DC is 0, but in the array is it 1 label for Cust 1 is 1, but in array it is 2, because DC is 1 T4+1 is the index for the DC-this is the "from" point TS+1 is the index for the DC-this is the "to" point so for the first stop "9", it looks up the matrix ID 10 Route To Travel Distance 97 INDEX(SC$4:SPS17,T4+1,T5+1) 13 INDEX(array, row-num, [column num All Solver engines in the Premium Solver Platform supports this new type of constraint. The Branch & Bound process used by the LP/Quadratic and GRG nonlinear Solvers is extended to handle "alldifferent" constraints as a native type, and the hybrid Evolutionary/Classical Solver implements these constraints using mutation and crossover operators for permutations." http://www.solver.com/premium-solver-platform-excel-alldifferent