In integrated circuit (IC) design, floorplanning is an important stage in obtaining the floorplan of the circuit to be designed. Floorplanning determines the performance, size, yield, and reliability of very large-scale integration circuit (VLSI) ICs. The results obtained in this step are necessary for the subsequent continuous processes of chip design. From a computational perspective, VLSI floorplanning is an NP-hard problem, making it difficult to be efficiently solved by classical optimization techniques. In this paper, we propose a deep reinforcement learning floorplanning algorithm based on sequence pairs (SP) to address the placement problem. Reinforcement learning utilizes an agent to explore the search space in sequence pairs to find the optimal solution. Experimental results on the international standard test circuit benchmarks, MCNC and GSRC, demonstrate that the proposed deep reinforcement learning floorplanning algorithm based on sequence pairs can produce a superior solution.

Reservoir Computing (RC) is a network architecture inspired by biological neural systems that maps time-dimensional input features to a high-dimensional space for computation. The key to hardware implementation of the RC system is whether sufficient reservoir states can be generated. In this paper, a laboratory-prepared zinc oxide (ZnO) memristor is reported and modeled. The device is found to have nonlinear dynamic responses and characteristics of simulating neurosynaptic long-term potentiation (LTP) and long-term depression (LTD). Based on this, a novel two-level RC structure based on the ZnO memristor is proposed. Novel synaptic encoding is used to maintain stress activity based on the characteristics of after-discharge and proneness to fatigue during synaptic transmission. This greatly alleviates the limitations of the self-attenuating characteristic reservoir of the duration and interval of the input signal. This makes the reservoir, in combination with a fully connected neural network, an ideal system for time series classification. The experimental results show that the recognition rate for the complete MNIST dataset is 95.08% when 35 neurons are present as hidden layers while achieving low training consumption.