Search

In this paper, we will discuss the issue of rostering jobs of cabin crew attendants at KLM.Generated schedules get easily disrupted by events such as illness of an employee. Obviously, reservepeople have to be kept ‘on duty’ to resolve such disruptions. A lot of reserve crew requiresmore employees, but too few results in so-called secondary disruptions, which are particularlyinconvenient for both the crew members and the planners. In this research we will discuss severalmodifications of the reserve scheduling policy that have a potential to reduce the number of secondarydisruptions, and therefore to improve the performance of the scheduling process.

In this paper, we will discuss the issue of rostering jobs of cabin crew attendants at KLM.Generated schedules get easily disrupted by events such as illness of an employee. Obviously, reservepeople have to be kept ‘on duty’ to resolve such disruptions. A lot of reserve crew requiresmore employees, but too few results in so-called secondary disruptions, which are particularlyinconvenient for both the crew members and the planners. In this research we will discuss severalmodifications of the reserve scheduling policy that have a potential to reduce the number of secondarydisruptions, and therefore to improve the performance of the scheduling process.

Splitting a large software system into smaller and more manageable units has become an important problem for many organizations. The basic structure of a software system is given by a directed graph with vertices representing the programs of the system and arcs representing calls from one program to another. Generating a good partitioning into smaller modules becomes a minimization problem for the number of programs being called by external programs. First, we formulate an equivalent integer linear programming problem with 0–1 variables. Theoretically, with this approach the problem can be solved to optimality, but this becomes very costly with increasing size of the software system. Second, we formulate the problem as a hypergraph partitioning problem. This is a heuristic method using a multilevel strategy, but it turns out to be very fast and to deliver solutions that are close to optimal.

Splitting a large software system into smaller and more manageable units has become an important problem for many organizations. The basic structure of a software system is given by a directed graph with vertices representing the programs of the system and arcs representing calls from one program to another. Generating a good partitioning into smaller modules becomes a minimization problem for the number of programs being called by external programs. First, we formulate an equivalent integer linear programming problem with 0–1 variables. Theoretically, with this approach the problem can be solved to optimality, but this becomes very costly with increasing size of the software system. Second, we formulate the problem as a hypergraph partitioning problem. This is a heuristic method using a multilevel strategy, but it turns out to be very fast and to deliver solutions that are close to optimal.