### Table 3: The initial candidate solution.

"... In PAGE 12: ... (See, for instance, the discussion by [15].) The search starts with the initial candidate solution described in Table3 . To generate neighboring solutions we use the perturbation operators described in Table 4.... ..."

### Table 3: The initial candidate solution.

"... In PAGE 12: ... #28See, for instance, the discussion by #5B15#5D.#29 The search starts with the initial candidate solution described in Table3 . To generate neighboring solutions we use the perturbation operators described in Table 4.... In PAGE 15: ...Allowing 2#25-6#25 for #0Cller gives a range of 42-44 pages, so we ran three sets of experiments in which ads were distributed initially over 42, 43, and 44 pages, respectively #28see Table3 #29. 3 Figures 4 through 6 plot objective-function scores #28using the objective function from Table 2#29 against the number of iterations #28one candidate solution is formulated and evaluated per iteration#29 for 20 runs of the algorithm, all seeded with di#0Berent random numbers.... ..."

### Table 3: The initial candidate solution.

in Harvard

1996

"... In PAGE 14: ... #28See, for instance, the discussion by #5B15#5D.#29 The search starts with the initial candidate solution described in Table3 . To generate neighboring solutions we use the perturbation operators described in Table 4.... In PAGE 17: ...Allowing 2#25-6#25 for #0Cller gives a range of 42-44 pages, so we ran three sets of experiments in which ads were distributed initially over 42, 43, and 44 pages, respectively #28see Table3 #29. 3 Figures 4 through 6 plot objective-function scores #28using the objective function from Table 2#29 against the number of iterations #28one candidate solution is formulated and evaluated per iteration#29 for 20 runs of the algorithm, all seeded with di#0Berent random numbers.... ..."

### Table 3. The initial candidate solution.

"... In PAGE 9: ...iminishing probability if worse. When convergence is reached, the process is terminated. Figure 3 contains a pseudocode description of our particular variant of simulated annealing. The search starts with the initial candidate solution described in Table3 . To generate neighboring solutions we use the perturbation operators described in Table 4.... In PAGE 12: ... Excluding filler, the ads and text in our test problem take up 40,8 pages. Allowing 2%-6% for filler gives a range of 42-44 pages, so we ran three sets of experiments in which ads were distributed initially over 42, 43, and 44 pages, respectively (see Table3 ) 3 . Figures 4 through 6 plot objective-function scores (using the objective function from Table 2) against the number of iterations (one candidate solution is formulated and evaluated per iteration) for 20 runs of the algorithm, all seeded with different random numbers.... ..."

### Table 3: Sample candidate solutions for simple example.

"... In PAGE 10: ... We evaluated all combinations of n; m 2 [1::16] for this exam- ple and our default model parameters. Table3 summarizes some of the candidate combinations. From left to right, the table lists the redundancy con guration, the number of disks, its throughput, its availability, the energy conservation technique, and the average power consumption.... ..."

### Table 4: Costs statistics for candidate solutions to the domestic problem

2003

"... In PAGE 19: ... Recall that the SAA method produces a number of candi- date solutions (at most M unique solutions). In Table4 , we compare statistics of the (uncertain) total cost for the domestic problem corresponding to the mean-value problem solution (denoted by yMV P ) to that of three candidate solutions (denoted by y1, y2, and y3 respectively) identified by solving M(= 20) SAA problem instances with N = 20. The candidate solutions were chosen as the three best solutions based upon their objective function value and optimality gap estimates as provided by the SAA method.... In PAGE 19: ... As before, the total cost statistics for each solution are computed using a sample size of Nprime = 1000. The last two rows of Table4 displays the estimated optimality gap and the standard deviation of the gap estimate (computed using (3.6) and (3.... ..."

### Table 1: Sample instance of Number Partitioning with two candidate solutions. In this

1994

"... In PAGE 8: ...optimal #28 Table1 #29. In pseudocode, the KK procedure is: for i =1to n , 1 j #20 the index of the largest elementofA k #20 the index of the second largest elementofA a j #20 a j , a k a k #20 0 end for When the procedure ends, at least n , 1 of the elements of A will be equal to zero, and the remaining element will be the value of u.... ..."

### Table 3 Candidate solutions: list of possible diagnosis, causes and control actions List of Diagnosis ((3) in Fig. 2)

"... In PAGE 7: ... Thus, in the bulking example, these goals correspond firstly to the definition of the prob- lem, secondly to the cause identification, and finally to the determination of the best plan to solve the problem. In the left column of Table3 the list of the possible diagnosis that the system can offer and that correspond to the Candidate Solutions in Fig. 2 is shown.... In PAGE 7: ...g., physical scum or foam removal, other concrete actuations), among those shown in the third row of Table3 . All of them are thus Candidate Solutions, but depending on the goal of the process (represented in different cycles in the model) one or another knowledge base (diagnosis, causes or actions) will be used.... ..."

### Table 1: Candidate solutions of 6y + 8z 0 (mod 13) tested by CBA y

"... In PAGE 3: ... As an example, let us take the equation 6 y + 8 z = 13 x. The remainders of 6 y + 8 z modulo 13 are shown in Table1 . Each null entry corresponds to the projection of a solution of the equation.... In PAGE 4: ...2 Basic Slopes Algorithm One of the drawbacks of CBA is that many of the candidate solutions failed the test for local minimality. For instance, in Table1 only 5 out of 14 candidate 2 solutions are (locally) minimal. The Slopes Algorithm also applies to Equation (2) but it improves CBA in that it moves from a m-solution to the next m-solution by nding the spacings between consecutive m-solutions.... ..."

Cited by 1