In this paper we present a genetic algorithm-based approach towards designing self-assembling objects comprised of square smart blocks. Each edge of each block can have one of three polarities (+1, -1 or 0) which defines how blocks stick together - opposite polarities attract, like polarities repel, and a 0 face neither attracts nor repels. In addition to this property, the block contains an internal state machine which can change the polarity of any number of its sides following the detection of an "event" (for example, two blocks sticking or unsticking). The aim of this work is to evolve block parameters and rule sets of the state machine which allow the self-assembly of desired basic structures that can be used as primitive building blocks for the assembly of more complicated objects. We detail a genetic algorithm-based approach that can be used to evolve the rule sets of interaction for a number of interacting blocks, so that the final shape or states of a structure formed by the blocks can approximate some target shapes or satisfy some global goals. We have assumed a list of simple identical properties for each block, and observed that a great diversity of complex structures can be achieved. © Springer-Verlag Berlin Heidelberg 2004.
CITATION STYLE
Guo, Y., Poulton, G., Valencia, P., & James, G. (2004). Designing self-assembly for 2-dimensional building blocks. In Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science) (Vol. 2977, pp. 75–89). Springer Verlag. https://doi.org/10.1007/978-3-540-24701-2_6
Mendeley helps you to discover research relevant for your work.