2 | | - ~~The ReferenceSetUpdateMethod could now use the CalculateCrowdSimilarity method added with r8304. I don't know if that makes the code in ReferenceSetUpdateMethod more complicated, but if not we should use this method because we have introduced it for exactly that case. |
3 | | - ~~ISimilarityCalculator should contain a documentation that explains what the range and the meaning of the returned value of CalculateIndividualSimilarity(..) is. |
4 | | - ~~TSPSimilarityCalculator should check for the different Permutation types. I think you should implement this in the same way as the PermutationEqualityComparer does. |
5 | | - ~~SimilarityCalculators should check that the length of the individuals is equal. |
6 | | - ~~Rename Target in SimilarityCalculator to VariableName. The Target parameter property should not be located in ISimilarityCalculator but in the problem-specific implementations. |
7 | | - ~~The same holds for the IImprovementOperator. The reason is that there could exist multiple targets when using an encoding that contains more than one component representing a solution. |
8 | | - ~~The comment in line 414 in ScatterSearch.cs has a typo. Maybe you should also add a comment that states that we have to think about the way we handle the target parameter again if we have an encoding where individuals consist of more than one target. |
9 | | - ~~UpdateSimilarityCalculators() in ScatterSearch.cs should use similarity instead of diversity for the naming of local variables. |
10 | | - ~~There should be an empty view for ISimiliarityCalculator so that HL doesn't display the "No view available" in the GUI. |
11 | | - ~~If you have based your improvers on code from a book or internet source, please add a comment with the link or reference. |
12 | | - ~~This should also be done for the Path Relinkers.~~ If you have a look at the crossover operators, there are always the paper or book references included in the description of the operator (ItemDescription) on which the implementation is based on. |
| 2 | - The ReferenceSetUpdateMethod could now use the CalculateCrowdSimilarity method added with r8304. I don't know if that makes the code in ReferenceSetUpdateMethod more complicated, but if not we should use this method because we have introduced it for exactly that case. |
| 3 | - ISimilarityCalculator should contain a documentation that explains what the range and the meaning of the returned value of CalculateIndividualSimilarity(..) is. |
| 4 | - TSPSimilarityCalculator should check for the different Permutation types. I think you should implement this in the same way as the PermutationEqualityComparer does. |
| 5 | - SimilarityCalculators should check that the length of the individuals is equal. |
| 6 | - Rename Target in SimilarityCalculator to VariableName. The Target parameter property should not be located in ISimilarityCalculator but in the problem-specific implementations. |
| 7 | - The same holds for the IImprovementOperator. The reason is that there could exist multiple targets when using an encoding that contains more than one component representing a solution. |
| 8 | - The comment in line 414 in ScatterSearch.cs has a typo. Maybe you should also add a comment that states that we have to think about the way we handle the target parameter again if we have an encoding where individuals consist of more than one target. |
| 9 | - UpdateSimilarityCalculators() in ScatterSearch.cs should use similarity instead of diversity for the naming of local variables. |
| 10 | - There should be an empty view for ISimiliarityCalculator so that HL doesn't display the "No view available" in the GUI. |
| 11 | - If you have based your improvers on code from a book or internet source, please add a comment with the link or reference. |
| 12 | - This should also be done for the Path Relinkers. If you have a look at the crossover operators, there are always the paper or book references included in the description of the operator (ItemDescription) on which the implementation is based on. |