A knowledge-based system to schedule multi-skilled labor with variable demand
The problem is represented by integer-programming model which is mainly considered here as a criterion for evaluating the efficiency of the proposed knowledge-based system in terms of the number of surplus workdays. Also, the proposed technique has the advantage of distributing the surplus workdays evenly.
thank u for comment. Yes, there are other techniques for solving this problem such as GA, ANN, tabu search, and more . they are good, but i think that KBS is an easy and efficiect way for scheduling labor.
in my system, i gurrantee that the soluion is optimum, because the mathematical equations of the ILP are translated into rules which are satisfied by the obtained solution.
i did a full survey for this problem, send me ur mail, today or tomorrow, if u want to take a look , my mail enhany75@yahoo.com.
What exactly do mean by variable demand? Since you fix the schedule for a whole week and if an urgent job comes, you will not be able to take it. So how flexible is your system.
Let us assume you are allowed 1 supervisor per shift, and the supervisor is capable of performing all the duties of his subordinates according to assumption #6. How would the system take care that the supervisor is not assigned an operator's job?
You mention KBS is easy and efficient way for scheduling labour compared to other AI techniques. What is the criteria for comparison?
Your communication is really interesting, since I think workforce scheduling is a more and more imortant issue in companies. Nevertheless, my opinion is that the research studies on the subject either start from "machine scheduling", and the workforce is poorly taken into account (mainly, the problem is to "allocate" an operator to each machine)à of from "staff/tour scheduling" like your's, and in this case basic concepts of production management like manufacturing orders, machines, routings or bills of materials are not really taken into account. In your model or instance, "tasks" must be achieved but these tasks do not have precedence constraints, due dates, raw material constraints which would allow to say that you plan the workforce in a manufacturing context. Am I right, and can you deal with theses constraints ?
Why did you choose to use Prolog? Wouldn't it be easyer with an expert system shell?
why do you choose knowledge-based system? what is the advantage? And what is the size of the knowledge base? Thanks.
May be we can also ask a question that is related to previous one. What is the matter of efficiency of this method over the others? And What is the defficiencies of this method? Many Thanks
I used Prolog because that is the most suitable langauge for this problem and most easer than expert system
using knowledge-based system to share and depend on the previous cases. The size of it can be increase with the user according to the cases applied.
Thanks
I can see that basically you've produced a sort of 'expert system' for labour scheduling using Prolog. Do you know of any other artificial intelligence techniques that have been applied to the problem of labour scheduling (e.g. ANNs GAs) and how sucessful they've been?