Advances in design and implementation of optimization software

Developing optimization software that is capable of solving large and complex real-life problems is a huge effort. It is based on a deep knowledge of four areas: theory of optimization algorithms, relevant results of computer science, principles of software engineering, and the state-of-the-art in computer technology. The paper highlights the diverse requirements of optimization software and discusses the ingredients needed to fulfil those requirements. In addition to reviewing the current hardware/software environment it hints at how recent hardware developments can influence future optimization software. After a survey of computationally successful techniques for continuous optimization, it also outlines the perspective offered by parallel computing, and stresses the importance of optimization modeling systems. Being a survey paper, it includes many references to both give due credit to results in the field of optimization software, and to help readers obtain more detailed information on issues of interest.