Talk:Shifting bottleneck heuristic

Making this more readable

Wow, that is intense. I'll have to read it a few more times to work out what you are saying and then suggest some way of making it simpler to understand. Its not the words, they're fine i think its the intensity of meaning or something. What is your relationship to this Heuristic ? Do you use it ? A lot ? Where did you get it from ? Do you know of lots of others who use it ?

Thinking about the general audience we should probably start on the intro and make it easy to read. So start by not introducing new words and meanings unless absolutely necessary. We probably need to start with the word 'makespan' in the intro. How about something like "minimise the overall time taken to complete a set of multi-machine jobs where machine order is pre-set for each job" or something.

BTW way to the welcome to the wikipaedia. Facius 11:46, 18 October 2007 (UTC)

Okay, i'll try to clarify a few things though the article. Obviously take them out if they are wrong or not what you meant. Facius 15:48, 22 October 2007 (UTC)

The inclusion of diagrams is a great step, but i have to admit that i cant understand how they are created or confirm what it may mean. I have tried to modify teh first section to be more readable and clear to me but ran out of confidence when i found i couldnt calculate the makespan myself. Maybe correct what i did and continue in that vein for further sections. Facius 16:13, 22 October 2007 (UTC)

The derivation of dij is not at all clear to me and so that whole step remains a mystery rather than an explanation to me. Facius 13:57, 26 October 2007 (UTC)

If I am not mistaking, the word "lateness" is never defined. Is it some function of dij, pij and rij ? The article wants to be saying how it is "determined", but this determination is rather implicit and not understandable if one does not know what lateness is.

Yes, oddly enough the referenced paper also does not define "lateness". I don't know if one of the authors of that paper is the author of this Wikepedia article too. Also I'm afraid I don't understand the definition of "due date". Subtracting the processing times of the preceeding tasks does not give you anything useful. It certainly does not give you the start time of a task itself since that may be delayed. AFAIK, the "due date" is simply a given, unless you specifically determine a modified due date, because you can either finish a task earlier than required or you are going to be late. So what I would like to see is:

• definition of due date as a problem given
• define completion time or finish time as the start time of a task + its duration
• define lateness as completion time minus due date

If there are no objections, I'm going to go ahead and make those changes

128.83.120.9 (talk) 20:31, 13 December 2007 (UTC)

Since I need to understand that algorithm now, I might come back and make the article clear when I am done.--129.88.43.7 (talk) 17:22, 20 November 2007 (UTC)