A Genetic Algorithm for Channel Routing in VLSI Circuits

A new genetic algoriihm for channel routing in $he physical design process of L'LSI circtiis i5 prestsfed. The algorithm is bused on a problem specific represen-iadion scheme and problem speciJc genetic operators. The geneiic encoding and our genelic operators are descrabed in defoii. The performance of.the algoriihnt is tested on difleerrnl benchmarks and it is shown that the results obdained using the proposed algorithm are either qttalatatiuely simo'lar t o o r better than the best published results. In the physical design procm of very large scale integrated (VLSI) circuits the logical structure of a circuit is transformed into its physical layout. Detailed routing is one of the tasks in this proces. A detailed router connects pins of signal nets in a rectangular region under a set of routing constraints, such as the number of layers, the minimal space between wires and the minimum wire width. The quality of this detailed routing has a strong influence on the perfoxm&nce and nroduction costs of the circuit. The detailed routing in a rectangular region with pins exclusively located on the upper or lower boundary of the routing region is called channel routing. Channel routing is one of the most commonly occurring routing problems in VLSI circuits. A simple example of a channel routing problem and a possible routing solution is shown in Figure 1. The channel routing problem is NP-complete I331 and therefore, there is no known deterministic alge rithm to solve it in a polynomial time. Hence, although many different algorithms have been proposed the problem of finding the globally optimized solution for channel routing is still open. Figure 1 : An example of a channel routing problem (a) and a p w i b l e routing solution (b). Solid lines represent interconilections on one layer, the poly Layer; and dashed lines represent interconnections on the other layer, the metal layer. New approaches are necessary to solve this problem. The evolution process in nature optimizes, for example, the fitness of an individual in its environment and thus, can be used as a strategy for mathematical optimization. Genetic algorithms are a new cIass of heuristic search methods based on the biological evolution model. During the last few years, genetic algorithms have been applied more and more s u c c m fully to find good heuristic solutjons to NP-complete optimization problems [13], 114. The strength of a genetic algorithm results from …