feat(coordinator-api): enhance reinforcement learning service with PyTorch-based PPO, SAC, and Rainbow DQN implementations

- Add PyTorch neural network implementations for PPO, SAC, and Rainbow DQN agents with GPU acceleration
- Implement PPOAgent with actor-critic architecture, clip ratio, and entropy regularization
- Implement SACAgent with separate actor and dual Q-function networks for continuous action spaces
- Implement RainbowDQNAgent with dueling architecture and distributional RL (51 atoms

apps/coordinator-api/systemd/aitbc-advanced-ai.service

@@ -0,0 +1,38 @@
+[Unit]
+Description=AITBC Advanced AI Service - Enhanced AI Capabilities
+After=network.target
+Wants=network.target
+
+[Service]
+Type=simple
+User=aitbc
+Group=aitbc
+WorkingDirectory=/opt/aitbc/apps/coordinator-api
+Environment=PATH=/opt/aitbc/.venv/bin
+Environment=PYTHONPATH=/opt/aitbc/apps/coordinator-api/src
+ExecStart=/opt/aitbc/.venv/bin/python -m app.services.advanced_ai_service
+ExecReload=/bin/kill -HUP $MAINPID
+Restart=always
+RestartSec=10
+StandardOutput=journal
+StandardError=journal
+SyslogIdentifier=aitbc-advanced-ai
+
+# Security settings
+NoNewPrivileges=true
+PrivateTmp=true
+ProtectSystem=strict
+ProtectHome=true
+ReadWritePaths=/opt/aitbc/logs /opt/aitbc/data
+
+# Resource limits
+LimitNOFILE=65536
+LimitNPROC=4096
+
+# GPU access (if available)
+DeviceAllow=/dev/nvidia0 rw
+DeviceAllow=/dev/nvidiactl rw
+DeviceAllow=/dev/nvidia-uvm rw
+
+[Install]
+WantedBy=multi-user.target