Computational Experimentation

In this chapter, we provide an overview of fundamental issues pertaining to performing computational experimentsComputational experiments in stochastic programming (SP). In addition to theory, models, and algorithms, implementation and application of the models and algorithms is also important. Implementing (coding) the models and algorithms on the computer requires computational experimentation. Therefore, it is fitting to end this book with a chapter on computational experimentation. We begin by reviewing problem data standard input formats for SP in Sect. 10.2. Because the input data formats involve using sparse matrices, we provide a review of sparse matrix formats in Sect. 10.3. We discuss program design for algorithm implementation and testing in Sect. 10.4 and end the chapter with a review of empirical analysis, methods of analysis, test problems, and reporting computational results in Sect. 10.5.