Faster-than-Nyquist (FTN) signaling is a potential scheme for the sixth generation (6G) communication system to improve the spectral efficiency (SE). In this paper

we propose a joint optimization algorithm of precoding and constellation shaping for FTN signaling

which is based on simulated optimization via the bare bones particle swarm optimization (BBPSO). The information-theoretical analysis and simulated error performance show that the proposed method is efficient

which can get a significant improvement in terms of average mutual information (AMI) and bit error rate (BER) performance. The simulated BER results verify the theoretical AMI analysis. Compared with the conventional regular 16QAM FTN scheme

when BER is at $10^{-5}$

the joint optimized scheme can obtain 0.50 dB and 0.60 dB performance gain with SE at 3.077 bits/s/Hz and 3.282 bits/s/Hz

respectively. Therefore

the proposed scheme is reliable

and thus suitable for the 6G communication.