From: G Bilchev Organization: University of Plymouth To: neum@cma.univie.ac.at, G.Bilchev@plymouth.ac.uk Date: Wed, 6 Mar 1996 22:49:35 GMT function: result: No. function calls: PERM(4,50) 1.7e-07 682 PERM(4,0.5) 6.3e-07 1967 PERM(4,0.05) 7.3e-07 1259 PERM(4,0.005) 5.4e-07 3562 PERM(4,0.0005) 7.7e-07 1301 PERM(4,0.00005) 7.6e-07 1329 PERM0(4,10) 2.6e-14 538 PERM0(4,1) 6.7e-13 596 PERM0(4,0.1) 1.6e-13 582 PERM0(4,0.01) 2.5e-13 607 TRID(4) -16 887 TRID(5) -29.99 1487 TRID(7) -76.99 2979 TRID(10) -209.95 6282 TRID(15) -662.63 14917 The standard bounds were used. POWERSUM(8,18,44,114) with bounds [0,4]^4 objective function value: 0.00057 at:(1.52092, 2.39422, 2.93307, 1.15462) for 33569 function calls. Target (1,2,2,3) is not reached. [f= 7.e-07 costs 33 000 000 function evaluations] The inductive search method used is deterministic, using function values only. See http://solon.cma.univie.ac.at/~neum/glopt/test_results.html#bilchev %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% From: G Bilchev Organization: University of Plymouth To: Arnold Neumaier Date: Fri, 22 Mar 1996 18:15:43 GMT I have increased the accuracy requirements within the local search. Here are updated results on the test problems: function value target No of calls PERM(4,50) 0 (1,2,3,4) 228 PERM(4,0.5) 0 (1,2,3,4) 208 PERM(4,0.005) 0 (1,2,3,4) 200 PERM(4,0) 0 (1,2,3,4) 201 PERM0(4,10) 6.7e-15 (1,0,5,0.333...,0.25) 36 PERM0(4,0.1) 1.0e-15 the same 36 PERM0(4,0.001) 9.89e-16 the same 36 PERM(4,0) the same the same 36 TRID(5) -30 as in the target 2952 TRID(10) -210 as in the target 21859 TRID(15) -665 as in the target 80373 TRID(20) -1520 as in the target 209645 TRID(100) 33000000 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%