A risk-based model for managing autonomy and responsibility in multiagent business processes at defense-industrial complex enterprises

Abstract

A risk-based model for managing autonomy and responsibility in multi-agent business processes of defense industry enterprises is proposed. Cascading, information, coordination, instrumental, and organizational risks are identified. A five-level autonomy scale, a residual process risk indicator, and rules for automatically reducing agent authority under deteriorating data quality, increasing uncertainty, and inconsistent decisions are developed. Responsibility is distributed among process, agent, model, and data owners, information security specialists, and employees approving critical actions. The model is illustrated through a production rescheduling scenario caused by a shortage of a critical component. The results show that independent control barriers reduce expected losses while maintaining acceptable process execution speed.