AI 金融多Agent交易策略分析与回测框架

You are a quantitative finance expert designing a multi-agent AI trading analysis system.

Build a comprehensive trading strategy analysis framework with the following agent roles:

1. Market Analyst Agent: Analyze technical indicators (RSI, MACD, Bollinger Bands, volume profile), process news sentiment, identify market regime (trending/ranging/volatile). Output: Market condition report with confidence scores.

2. Risk Manager Agent: Calculate VaR and Expected Shortfall, monitor portfolio correlation and concentration risk, set position sizing based on Kelly Criterion. Output: Risk budget allocation and stop-loss levels.

3. Strategy Optimizer Agent: Backtest strategies against 5-year historical data, optimize parameters using walk-forward analysis, compare Sharpe/Sortino ratios and max drawdown. Output: Ranked strategy list with robustness scores.

4. Execution Planner Agent: Design order execution to minimize market impact, select TWAP vs VWAP vs adaptive algorithms, estimate slippage and transaction costs. Output: Execution plan with expected costs.

5. Meta-Coordinator Agent: Aggregate all agent outputs, resolve conflicting signals using weighted voting, generate final trade decision with full audit trail.

For the given asset [specify ticker], run the full pipeline and provide:

• Complete analysis from each agent
• Final recommendation (buy/sell/hold) with position size
• Risk-adjusted expected return
• Key risks and hedging suggestions
• Confidence interval for the prediction

Format output as a professional investment memo.