X-structure steiner minimum tree routing method based on adaptive pso and hybrid transformation strategy

Abstract

The invention relates to an X-structure Steiner minimum tree algorithm based on an adaptive PSO and a mixed transformation strategy. The X-structure Steiner minimum tree algorithm based on the adaptive PSO and the mixed transformation strategy comprises the following steps: reading and sequencing test circuit data; initializing all parameters of the algorithm, and randomly generating an initial population; calculating a fitness value of each particle, and selecting the particle with the minimum fitness value as the global optimum of the population, and setting every particle in the first generation as the historical optimum; taking the mixed transformation strategy as basic operation of particle updating; updating the positions and the speeds of the particles; recalculating the fitness values of the particles, if the current fitness values of the particles are smaller than the historical optimum, setting the updated particles as the historical optimum; if the fitness values of the updated particles are smaller than the fitness value of the global optimal particle of the population, setting the updated particles as the global optimum of the population; and if the maximum iterative times of the algorithm are achieved, terminating the algorithm, and if the maximum iterative times of the algorithm are not achieved, returning to updating operation. By the X-structure Steiner minimum tree algorithm based on the adaptive PSO and the mixed transformation strategy, optimization of a line length is used as a target, and finally, the line length, which is the important target, is optimized.

Description

technical field [0001] The invention relates to a X-structure Steiner minimum tree wiring method based on adaptive PSO and hybrid conversion strategy Background technique [0002] General layout is a very important step in the physical design of very large scale integration (VLSI). The Steiner Minimal Tree (SMT) problem is to introduce some additional points (Steiner points) on the basis of a given set of pins to find a minimum-cost wiring tree connecting a given set of pins. Because the SMT model is the best connection model of multi-terminal nets in the overall wiring of VLSI, the construction of SMT is a key link in the wiring of VLSI. [0003] In recent years, integrated circuits have developed rapidly in terms of design scale and manufacturing process, which brings new challenges to electronic design automation. The problems faced by the overall routing stage in the physical design are also becoming more and more complex. The research work on the overall routing probl...

AI Technical Summary

Problems solved by technology

[0006] Constructing a non-Manhattan structure Steiner tree is an NP-hard problem. The above research work on solving the non-Manhattan structure Steiner tree is based on the exact algorithm and the traditional heuristic algorithm, and the time complexity of the exact algorithm increases exponentially with the scale of the problem. Most of the traditional heuristic algorithms are based on greedy strategies, which are easy to fall into local extremum

Moreover, when constructing the Steiner tree, the two did not make full use of the geometric properties of the non-Manhattan structure, so the quality of the Steiner tree cannot be guaranteed, and the main optimization goal is focused on line length optimization, and few suitable methods for topology optimization are provided, resulting in time efficiency. , line length shortening, etc. are still unsatisfactory

Images