Lotsizing and scheduling for production planning (Q1327721)

The book gives the reader an insight into the variety and complexity of practical lotsizing and scheduling problems and solution methods. It demonstrates properties and difficulties mainly using small instructive examples. A more detailed description of the content of the book is given below: (1) Lot sizing and scheduling models are introduced. First, a list of some different aspects for objectives is provided resulting in a classification of lot sizing and scheduling models. (2) Then, a simple model for a lot sizing problem is formulated, the disadvantages are demonstrated on an example and a more realistic extension of the model is provided. Both models consist of an integer linear program. Solution methods are provided for both models taking into account the extended model. (3) Further extensions are stated to treat more aspects of lotsizing models. These extensions are compared regarding the quality of the solutions, i.e. the variety of possible solutions. Therefore an overview is given for including further aspects to derive a richer set of feasible solutions. (4) Then sequencing aspects, which have been neglected in the previous sections, are considered, and further extensions are developed (e.g. buy or make, backordering and stockouts). (5) A stochastic algorithm is described and analyzed. This algorithm is based on dividing the parameter space into subspaces and choosing subspaces which provides ``good'' solutions by a sequential test method (checking quantiles of random solutions). (6) This stochastic algorithm is applied to the basic lotsizing problems introduced in (2). The algorithm, called ``backward add method'', is related to well-known local search methods. (7) Computational results are provided using the proposed algorithm. Numerical aspects are discussed and further promising extensions of the stochastic algorithm are stated.