Enumeration methods 1: Unterschied zwischen den Versionen
[unmarkierte Version] | [unmarkierte Version] |
Pvogt (Diskussion | Beiträge) (→Implicit complete enumeration) |
|||
Zeile 19: | Zeile 19: | ||
In case of large scale optimization problems the computing time for complete enumeration techniques often exceeds the economic reasonable time. Therefore a use of incomplete enumeration (heuristics) is advised. Incomplete enumeration techniques just search in a subset of the valid solution space. Hence the best generated solution is not necessarily a optimal solution. It is also possible, that a heuristic stops without a valid solution even if there is one. | In case of large scale optimization problems the computing time for complete enumeration techniques often exceeds the economic reasonable time. Therefore a use of incomplete enumeration (heuristics) is advised. Incomplete enumeration techniques just search in a subset of the valid solution space. Hence the best generated solution is not necessarily a optimal solution. It is also possible, that a heuristic stops without a valid solution even if there is one. | ||
− | == Example == | + | == Example: Traveling Salesman Problem (TSP) == |
+ | |||
+ | In the following we discuss one examples which we solve with the three enumeration methods mentioned before. The following table shows the cost a salesman has to travel from a location i to a location j. | ||
+ | |||
+ | Table 1: | ||
+ | [[Datei:Tabelle_1.png]] | ||
== Presentation of the problem == | == Presentation of the problem == |
Version vom 1. Juli 2013, 19:34 Uhr
There are two solution sets for difficult combinatorical optimization problems available: complete and incomplete enumeration methods. These solution sets also apply on integer or mixed-integer optimization problems.
Inhaltsverzeichnis
Theory
Complete enumeration
A complete enumeration method guarantees, that from all possible solutions the optimum is chosen. Algorithms that always provide the optimal solution are called exact methods.
Explicit complete enumeration
One possibility to solve a problem is explicit complete enumeration. The algorithm generates consecutively a enumeration tree with all possible solutions and memorizes the current best solution of the examined problem.
Implicit complete enumeration
Another possibility to exactly solve a combinatorical optimization problem is implicit complete enumeration. The algorithm consecutively cuts of all subsets of the solution space in which, under predetermined conditions, no optimal solution could be expected. The algorithm generates consecutively a enumeration tree of the remaining solution space.
Incomplete enumeration
In case of large scale optimization problems the computing time for complete enumeration techniques often exceeds the economic reasonable time. Therefore a use of incomplete enumeration (heuristics) is advised. Incomplete enumeration techniques just search in a subset of the valid solution space. Hence the best generated solution is not necessarily a optimal solution. It is also possible, that a heuristic stops without a valid solution even if there is one.
Example: Traveling Salesman Problem (TSP)
In the following we discuss one examples which we solve with the three enumeration methods mentioned before. The following table shows the cost a salesman has to travel from a location i to a location j.
Table 1:
Presentation of the problem
Solution process
Sources
Neumann, K.; Morlock, M.: Operations Research, 2.Auflage, Carl Hanser Verlag München Wien, 2002
Zimmermann, W.:Operations Research, Quantitative Methoden zur Entscheidungsvorbereitung, 6.Auflage, R.Oldenbourg Verlag München Wien, 1992