Progressively Interactive Evolutionary Multi-Objective Optimization ...
Progressively Interactive Evolutionary Multi-Objective Optimization ...
Progressively Interactive Evolutionary Multi-Objective Optimization ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Table 6. Distance of obtained solution from the most preferred solution, function evaluations, and<br />
the number of DM calls required by PI-NSGA-II for the five-objective modified DTLZ2 problem.<br />
ds = 0.01 ds = 0.1<br />
Minimum Median Maximum Minimum Median Maximum<br />
Accuracy 0.0112 0.0329 0.1210 0.0395 0.0884 0.2777<br />
Func. Evals. 20272 29298 37776 5083 6872 9919<br />
# of DM Calls 51 69 96 9 12 17<br />
6 Parametric Study<br />
Besides the usual parameters associated with an evolutionary algorithm, such as population<br />
size, crossover and mutation probabilities and indices, tournament size etc., in<br />
the proposed PI-NSGA-II we have introduced a few additional parameters which may<br />
effect the accuracy and number of DM calls. They are the number of generations between<br />
DM calls (τ), termination parameter (ds), maximum archive size (|A| max ), KKT<br />
error limit for terminating SQP algorithm in single-objective optimization used for the<br />
termination check, and the parameter ρ used in the ASF function optimization. Of these<br />
parameters, the first two have shown to have an effect on the chosen performance measures<br />
— accuracy, the number of overall function evaluations, and the number of DM<br />
calls.<br />
A parametric study for ds has not been done in this section as results for two different<br />
values of ds have already been presented in the previous section. The results show<br />
an expected behavior, that is, a strict ds provides higher accuracy and requires a larger<br />
number of DM calls and function evaluations, a relaxed ds provides lower accuracy and<br />
requires less number of DM calls and function evaluations.<br />
Thus, in this section, we study the effect of the parameter τ, while keeping ds =<br />
0.01 and the other PI-NSGA-II parameters identical to that mentioned in the previous<br />
section. Here, we use the two objective ZDT1 and three objective DTLZ2 test problems.<br />
6.1 Effect of Frequency of DM Calls (τ )<br />
We study the effect of τ by considering four different values: 2, 5, 10 and 20 generations.<br />
The parameter ds is kept fixed at 0.01. To investigate the dependence of the<br />
performance of the procedure on the initial population, in each case, we run PI-NSGA-<br />
II from 21 different initial random populations and plot the best, median and worst<br />
performance measures.<br />
We plot three different performance measures — accuracy, number of DM calls and<br />
number of function evaluations obtained for the modified ZDT1 problem in Figure 9.<br />
It is interesting to note that all three median performance measures are best for τ =<br />
5. A small value of τ means that DM calls are to be made more frequently. Clearly,<br />
this results in higher number of DM calls, as evident from the figure. Frequent DM<br />
calls result in more single-objective optimization runs for termination check, thereby<br />
increasing the number of overall function evaluations. On the other hand, a large value<br />
of τ captures too little information from the DM to focus the search near the most<br />
69