Seite - 185 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments
Bild der Seite - 185 -
Text der Seite - 185 -
6. Mutation:ApplyDecrementMutation to randomlychosenN/20solutions (Fig.
3), and apply IncrementMutation to otherN/20 solutions (Fig. 2). Then, go to
step2after incrementing t.N/20waschosenbasedon thedefaultmutation rate
0.1used inNSGA-II.
4.3. Mutation
In our algorithm, we designed twoMutation algorithms: one is to randomly insert one
PoI and the other is to randomly remove onePoI in a solution so that diverse solutions
are kept in terms of the number of PoIs.One of the algorithms is selected to use at the
probabilityof0.5.
IncrementMutation:As shown inFig. 2, first aPoI (PoI7 in the fuigure) is randomly
chosen from the set of all PoIs except for the ones already in the solution and a trans-
portation type leaving from the addedPoI (Car in thefigure) is randomly chosen.Next,
the insert position in a solution is randomly chosen, andboth the chosenPoI and trans-
portation typeare inserted.
DecrementMutation: This algorithm is applied only to a solution with lengthmore
than two. If the length ismore than two, as shown inFig. 3, the point to removePoI is
randomly chosen (PoI 21was chosen in the figure) and the PoI and the transportation
type leaving from thePoI (Car in thefigure) are removed.
Figure2. IncrementMutation Figure3. DecrementMutation
4.4. Crossover
We employ single-point crossover where a single cut point in each of two parents is
randomly selected and the left (right) part of the parent 1 and the right (left) part of
the parent 2 are swapped tomake newoffspring, as shown in Fig. 4.However, simply
using the single-point crossover forour algorithm, it is likely togenerate lethal (invalid)
solutionswhich include the same PoImultiple times. Hence, before concatenating the
divided parts of the parent solutions, we try to reduce generation rate of such lethal
solutionsbyrandomlyreplacingeitherof thePoIs includedintheleftpartofaparentand
rightpartofanotherparent (middleofFig.4).Forconvenience,whenwedivideaparent
solution into twoparts,we remove the incoming transportation type in the rightpart and
add a randomly selected transportation typewhen concatenating the parts as shown in
Fig.4. Y.Hiranoetal. /AMethod forGeneratingMultipleTourRoutes 185
Intelligent Environments 2019
Workshop Proceedings of the 15th International Conference on Intelligent Environments
- Titel
- Intelligent Environments 2019
- Untertitel
- Workshop Proceedings of the 15th International Conference on Intelligent Environments
- Autoren
- Andrés Muñoz
- Sofia Ouhbi
- Wolfgang Minker
- Loubna Echabbi
- Miguel Navarro-Cía
- Verlag
- IOS Press BV
- Datum
- 2019
- Sprache
- deutsch
- Lizenz
- CC BY-NC 4.0
- ISBN
- 978-1-61499-983-6
- Abmessungen
- 16.0 x 24.0 cm
- Seiten
- 416
- Kategorie
- Tagungsbände