f func_b: The number of functions where the DFPSO algorithm obtains significantly better results。
g func_w: The number of functions where the DFPSO algorithm obtains significantly worse results。
Table 5
Statistical results obtained by the Wilcoxon signed ranks test for pairwise comparisons based on 25 benchmark functions。
Comparison Rþa R— b p value Comparison Rþ R— p value
DFPSO versus PSO 275 50 0。0016 DFPSO versus PSOw 280。5 44。5 0。0011
DFPSO versus ALC-PSO 201。5 123。5 P 0:2 DFPSO versus SPSO2007 226。5 98。5 0。0951
DFPSO versus GPSO 269。5 55。5 0。0029 DFPSO versus SPSO2011 247 78 0。0219
DFPSO versus CLPSO 227。5 97。5 0。0826
a Rþ: The sum of ranks for the functions in which the first algorithm outperforms the second one。
b R— : The sum of ranks for the functions in which the second algorithm outperforms the first one。
5。Odor source localization
In this section, we will deal with the problem of odor source localization by using the proposed continuous-time FPSO algorithm, which consists of a decision algorithm pi ðkÞ and a finite-time cooperative control algorithm ui ðtÞ given by (17)。 In terms of the characteristics of the odor source localization problem, the decision algorithm has been developed in our pre- vious studies [23,24]。 In order to let the reader easily understand the proposed continuous-time FPSO algorithm, we will
briefly restate the decision algorithm。
5。1。The decision algorithm
The probable position of the odor source estimated by the ith robot based on wind information and concentration infor- mation can be calculated by
hi ðkÞ¼ c1^xi ðkÞþ c2 a ð Þþ a ð Þ
where cj ðj ¼ 1; 2Þ are weighted parameters to be selected to satisfy c1 þ c2 ¼ 1; ^xi ðkÞ is the position of the odor source esti-
mated by the ith robot based on wind information while robot based on concentration information。
a1 þa2 is the position of the odor source estimated by the ith
In terms of wind information [23,24], the dynamics model of the position of the odor source is directly given by
。 xs ðkÞ¼ xs ðk — 1Þ