DPSO based Octagonal Steiner Tree algorithm for VLSI routing

Abstract

The Octagonal Steiner Minimal Tree (OSMT) problem is an NP-hard problem, which is one of the key problems in non-Manhattan routing. Particle Swarm Optimization (PSO) has been proved to be an efficient intelligent algorithm for optimization designs. This paper presents an OSMT algorithm based on discrete PSO (DPSO), namely OSMT_DPSO, to optimize the wire length. In order to solve the problem of the slow convergence rate of PSO used for a high-dimensional space optimization, a self-adapting strategy that can adjust the learning factors, and combine with the crossover and mutation operators of Genetic Algorithm (GA) is proposed. The experimental results show that the proposed algorithm can efficiently provide the solution of OSMT problem with good quality. Moreover, the algorithm can obtain several topologies of OSMTs which is beneficial for optimizing congestion in VLSI global routing stage.

A PSO-based timing-driven Octilinear Steiner tree algorithm for VLSI routing considering bend reduction

“…In contrast to previous work (Liu et al 2012) which only takes into account the transformation of Steiner point, we introduce the edge-transformation strategy to improve the evolutionary process of the particles. And the novel strategy could simultaneously transform the Steiner points and edges of the routing tree.…”

“…As a result, the method with the encoding strategy can reduce nearly 20 % in the number of bends. However, the work (Liu et al 2012) did not consider the bend reduction without using the edge transformation. 3.…”

“…In order to simultaneously optimize the two objectives of wire length and radius which are often competing objectives, we design a multi-objective algorithm based on discrete particle swarm optimization (PSO) for constructing the timing-driven Octilinear routing tree. The experimental results show that the final trees achieve a major decrease in maximum source-to-sink delay and total source-tosink delay than the RSMTs and OSMTs which focus on the optimization of wire length and are respectively constructed in one literature ) and the other literature (Liu et al 2012). Moreover, based on the 0.18 µm technology (namely, tost IC technology), the final tree constructed by our algorithm could obtain 2.08 % smaller than the TOST constructed by Yan (2008) in wire length, 5.22 % smaller in radius, 4.93 % smaller in the maximum source-to-sink delay, 10.07 % smaller in the total source-to-sink delay and nearly 20 % smaller in the number of bends.…”

“…And the novel strategy could simultaneously transform the Steiner points and edges of the routing tree. Consequently, compared with OST (Liu et al 2012), the routing tree can decrease a certain amount of wire length on one hand. It is very necessary that the initial particles have the ability to transform to MRMCST by the introduction of edge-transformation strategy on the other hand.…”

“…Samanta et al (2006) proposed a heuristic method to constructing Hexagonal Steiner minimal tree. Since the SMT problem is an NPhard problem (Garey and Johnson 1997), some evolutionary algorithms, which have been shown to have good application prospects in solving NP-hard problems, were presented for solving RSMT problem (Arora and Mose 2009;Liu et al 2011) and OSMT problem (Arora and Mose 2009;Liu et al 2012).…”