1

Topic: Manufacture optimization

There is a set of technological operations. Each operation has following attributes 1) D (n) - what is the time (in minutes) it is required on operation performance n 2) T (n) - date and time, to which operation n should be fulfilled 3) S (n) - date and starting time of performance of operation n. Date and ending time of performance of operation n: F (n) = S (n) + D (n). 4) With (n) - an integer number from 0 to 100, characterizing parameters of adjustment of the technology equipment, necessary for the fulfilled operation n Operations are fulfilled one for another that is as soon as one operation ended at once begins another. If one after another there are operations with identical value of a C it positively affects technology equipment (the less the equipment is required to recustomize, the it serves longer, the probability of an error there is less at realignment). On an input the following data is known: 1) the Amount of technological operations N 2) {D (1), D (2), ..., D (N)} 3) {T (1), T (2), ..., T (N)} 4) {a C (1), a C (2), ..., a C (N)} 5) B - date and time when should begin the first operation. It is required to calculate {S (1), S (2), ..., S (N)}, thus that  (T (n) - F (n)) was minimum, and the amount of sequentially executable operations with the same value With was maximum. I will be glad to any helps to the task decision.

2

Re: Manufacture optimization

Hello, igor-booch, you wrote: IB> I Will be glad to any helps to the task decision. Cite the test input data. As it is a lot of operations N: tens, hundreds, hundred thousand? In target function  yours  a C (i). And then it minimize. Generally similar tasks are reduced to tasks with exponential growth in an exact variant, so to you  any greedy algorithms and heuristics.

3

Re: Manufacture optimization

_> As it is a lot of operations N: tens, hundreds, hundred thousand? Hundreds, duration of everyone approximately from 3 till 96 o'clock. _> generally similar tasks are reduced to tasks with exponential growth in an exact variant, so to you  any greedy algorithms and heuristics. Yes, it is possible to make stupidly a decision tree, but it is necessary to be able to cut branches by any criteria. There are ideas how to cut a branch?

4

Re: Manufacture optimization

Hello, igor-booch, you wrote: _>> As it is a lot of operations N: tens, hundreds, hundred thousand? IB> hundreds, duration of everyone approximately from 3 till 96 o'clock. _>> generally similar tasks are reduced to tasks with exponential growth in an exact variant, so to you  any greedy algorithms and heuristics. IB> yes, it is possible to make stupidly a decision tree, but it is necessary to be able to cut branches by any criteria. There are ideas how to cut a branch? For this purpose it is necessary, test given ideas how to cut appear cancers , at once.

5

Re: Manufacture optimization

Hello, igor-booch, you wrote: I Think here well the genetic algorithm approaches.

6

Re: Manufacture optimization

Hello, igor-booch, you wrote: IB> There is a set of technological operations.... IB> I will be glad to any helps to the task decision. And you any else algorithms, except tree creation know? Well for example "a method of the left end"? It when segments are laid down on horizontals. And when the next horizontal is released, the next segment is laid down on it. The method is used at trace of chips, but it can be used for equipment loading.