Abstract
The nesting problem, also known as irregular packing problem, belongs to the generic class of cutting and packing (C&P) problems. It differs from other 2-D C&P problems in the irregular shape of the pieces. This paper proposes a new mixed-integer model in which binary decision variables are associated with each discrete point of the board (a dot) and with each piece type. It is much more flexible than previously proposed formulations and solves to optimality larger instances of the nesting problem, at the cost of having its precision dependent on board discretization. To date no results have been published concerning optimal solutions for nesting problems with more than 7 pieces. We ran computational experiments on 45 problem instances with the new model, solving to optimality 34 instances with a total number of pieces ranging from 16 to 56, depending on the number of piece types, grid resolution and the size of the board. A strong advantage of the model is its insensitivity to piece and board geometry, making it easy to extend to more complex problems such as non-convex boards, possibly with defects. Additionally, the number of binary variables does not depend on the total number of pieces but on the number of piece types, making the model particularly suitable for problems with few piece types. The discrete nature of the model requires a trade-off between grid resolution and problem size, as the number of binary variables grows with the square of the selected grid resolution and with board size.

Abstract

Abstract

Abstract
Metadata production for research datasets is not a trivial problem. Standardized descriptors are convenient for interoperability, but each area requires specific descriptors in order to guarantee metadata comprehensiveness and accuracy. In this paper, we report on an ongoing research data management experience at the University of Porto (U. Porto), which led to the proposal of a domain-specific application profile. We presented two curation tools to a group of researchers from mechanical engineering, to help them manage and describe their datasets. After monitoring their interactions with the solutions and analyzing the needs of the group, we were able to select a subset of qualified Dublin Core (DC), as well as a set of complementary descriptors, to capture the main aspects of their experiments. The resulting application profile combines generic, standardized DC descriptors with descriptors from a different experimental standard, and introduces extra domain-specific ones. The profile has been validated by the researchers and is now being used in the description of their datasets. © DCMI 2013.

Abstract

Abstract