the up bounds do not work correctly

[MontyKL]

when the precision is one, ie. the int_mode is true, the optimized result will lager than the up bounds.

[MontyKL]

at GA.py line 154, when you the constriant, the attribute of has constriant should be set to TRUE, self.has_constriant = True

[guofei9987]

Thank you!
You are right, it is a bug. I have fixed it.

• if precision is integer:
• if Lind_raw is integer, which means the number of all possible value is $2^n$, no need to modify
• if Lind_raw is decimal, modify: make the ub bigger and add a constraint_ueq

The fixed version of scikit-opt will be released next version, and I‘d like to list you as contributor if you permit.

[guofei9987]

Thanks again! And I find more bugs in int_mode recently.
Besides, using penalty function to deal with upper bounds is a bad idea. It causes a lot of misusage and misunderstandings.
I have rebuilt the algorithm by another way in version 0.5.7.

[Jackwangnao]

郭老师您好！还想接着问一个整数模式的问题：
int_mode为True是指整数规划模式开启对吧？所以源码中self.int_mode_ = (self.precision % 1 == 0) & (Lind_raw % 1 != 0)这里的 Lind_raw 应该判断为 Lind_raw % 1 == 0 吧？

[guofei9987]

郭老师您好！还想接着问一个整数模式的问题：
int_mode为True是指整数规划模式开启对吧？所以源码中self.int_mode_ = (self.precision % 1 == 0) & (Lind_raw % 1 != 0)这里的 Lind_raw 应该判断为 Lind_raw % 1 == 0 吧？

这里代码可能有些歧义，这里的“整数规划模式”指的是需要特别处理才是整数的。如果问题本身不经过特别处理已经是实质上的整数规划，就没必要开启整数规划模式。