ABSTRACT

This PhD thesis deals with the analysis and modelling of the problems arising in the programming of cutting operations on structural steel profiles, and the proposal of several methodologies and algorithms based on heuristics techniques that solve them optimally. In particular, the thesis focuses on the following issues:

- Considers the specific problem arising in cutting structural steel beams in a national manufacturer. This study motivates and justifies all subsequent work, while providing a specific context in which applying the results obtained with the developed algorithms. 

- The Cutting Stock Problem is modelled and identified in the national manufacturer of metal profiles.

- An efficient methodology, based on the use of cutting patterns, is described in order to solve the Cutting Stock Problem so that all costumers´ demands are satisfied for a period of time. So far, we develop the following: a genetic algorithm that generates efficient cutting patterns (phase 1); a second genetic algorithm which solves in a first step the cutting stock problem by determining the frequencies of use of each pattern (phase 2); and four additional algorithms which improve the solution obtained in the previous phase (phase 3).

- In order to evaluate the proposed methodology, a problem generator is developed, so that from certain parameters of an instance the generator gives rise to specific test problems.

- Another genetic algorithm is proposed to solve the multi-objective Sequencing Problem of cutting patterns so that two objectives are accomplished simultaneously: the minimization of space requirements to pile work of in process orders.

- Finally, it is proposed a methodology to solve the Global Cutting and Sequencing Problem