everybody. I am working on a university timetable scheduler project. Mainly, I am using taboo search, but I want to ask:
In general search, you can explore all neighbors of the current state and then take the best state – according to a fitness or evaluation function, – but in such a project, generating all neighbors will make performance down, so is there any way that make me bypass such problem? For example, can I generate children for only one state and then benefit from this generation for all other states during the search process?
Please, if anyone has an expert in such algorithms, please tell me, because I have worked hard on such issues.