Abstract
This work describes a real-world industrial problem of production planning and cutting optimization of reels and sheets, occurring at a Portuguese paper mill. It will focus on a particular module of the global problem, which is concerned with the determination of the width combinations of the items involved in the planning process: the main goal consists in satisfying an order set of reels and sheets that must be cut from master reels. The width combination process will determine the quantity/weight of the master reels to be produced and their cutting patterns, in order to minimize waste, while satisfying production orders. A two-phase approach has been devised, naturally dependent on the technological process involved. Details of the models and solution methods are presented. Moreover some illustrative computational results are included.

Abstract
Nesting problems are particularly hard combinatorial problems. They involve the positioning of a set of small arbitrarily-shaped pieces on a large stretch of material, without overlapping them. The problem constraints are bidimensional in nature and have to be imposed on each pair of pieces. This all-to-all pattern results in a quadratic number of constraints. Constraint programming has been proven applicable to this category of problems, particularly in what concerns exploring them to optimality. But it is not easy to get effective propagation of the bidimensional constraints represented via finite-domain variables. It is also not easy to achieve incrementality in the search for an improved solution: an available bound on the solution is not effective until very late in the positioning process. In the sequel of work on positioning non-convex polygonal pieces using a CLP model, this work is aimed at improving the expressiveness of constraints for this kind of problems and the effectiveness of their resolution using global constraints. A global constraint "outside" for the non-overlapping constraints at the core of nesting problems has been developed using the constraint programming interface provided by Sicstus Prolog. The global constraint has been applied together with a specialized backtracking mechanism to the resolution of instances of the problem where optimization by Integer Programming techniques is not considered viable. The use of a global constraint for nesting problems is also regarded as a first step in the direction of integrating Integer Programming techniques within a Constraint Programming model.

Abstract
In this paper we describe random start local search and tabu search for solving a multi-item, multi-machine discrete lot sizing and scheduling problem with sequence dependent changeover costs. We present two construction heuristics with a random component; one of them is purely random and another is based on the linear programming relaxation of the mixed integer programming model. They are used to generate initial solutions for random start local search and tabu search. We also propose two ways of exploring the neighborhoods, one based on a random subset of the neighborhood, and another based on exploring the whole neighborhood. Construction and improvement methods were combined on random start local search and tabu search, leading to a total of eight different methods. We present results of extensive computer experiments for analyzing the performance of all methods and their comparison with branch-and-bound, and conclude with some remarks on the different approaches to the problem.

Abstract
In this paper, we introduce a GRASP for the solution of general linear integer problems. The strategy is based on the separation of the set of variables into the integer subset and the continuous subset. The integer variables are fixed by GRASP and replaced in the original linear problem. If the original. problem had continuous variables, it becomes a pure continuous problem, which can be solved by a linear program solver to determine the objective value corresponding to the fixed variables. If the original problem was a pure integer problem, simple algebraic manipulations can be used to determine the objective value that corresponds to the fixed variables. When we assign values to integer variables that lead to an impossible linear problem, the evaluation of the corresponding solution is given by the sum of infeasibilities, together with an infeasibility flag. We report results obtained for some standard benchmark problems, and compare them to those obtained by branch-and-bound and to those obtained by an evolutionary solver.

Abstract

Abstract
The commercial use of the Internet for service provision has deeply changed the environment where human-computer interaction takes place, as Web interfaces are now integrated in multi-platform service provision. This paper presents the results of a study of a multi-channel Portuguese bank, making use of both Marketing and HCI methods and concepts, to understand customer usage of the different service platforms. The study involved in-depth interviews, focus groups, a web survey and a telephone survey with bank customers. The study allowed the identification of the most important interaction experience requirements for this multi-platform service, and how they are influenced by user profiles and service characteristics. The results also show that Customer Experience Requirements (CERs) have a strong impact on customer choice and usage of the different service platforms and can be better captured with Essential Use Cases (EUCs), as they are technology independent. Designing a multi-platform service interaction should therefore start with a higher level of abstraction that allows a multi-platform, customer experience and essential use case perspectives. With this integrated approach, the Internet service can therefore be designed in order to best leverage its capabilities and its complementarity with the other service platforms.