PLACING COMPUTATIONS BY ADAPTIVE PROCEDURES

The present work treats the computation of heterogeneous layout config urations; distinct shapes as rectangles and triangles have to get plac ed into a target frame. In our design application the main restriction is the requirement that the objects must not overlap. Here, we furthe r constrain the problem to obey the following requirement: The user sh all be able to interact with the (semi-)automatic layout system in a w ay such that s/he may pick an object to place it in a subarea, offered by the system, in an arbitrary manner without the need to think about the placement of the other objects. (After an algorithm's termination a globally consistent layout should be guaranteed.) Thereby, the user still has degrees of freedom to finally arrange the objects. Such a r ealization shall also enhance the acceptance of the system by the user because entire solution classes, obtained by topological layout relat ions (instead of maintaining single co-ordinate points), are offered. The current work employs evolutionary computing techniques in order to get timely computations. This paper shows the use of interesting arti ficial intelligence techniques in the design area with close connectio ns to a class of combinatorial problems in operational research with a wide range of applications in business and industry. (C) 1997 Elsevie r Science Limited.