← Back to Home

🚨 Park Emergency Scenario Demo

DSL Multi-Agent Framework - Real Emergency Response Simulation

🎯 Goals · Roles · I/O

Goals

  • ⏳ Response time within SLA window
  • ⏰ All tasks meet deadlines

Roles & Agents

  • 🧑‍💼 Dispatcher: spawn, route
  • 🩺 Medics: route, contract
  • 🚁 UAV: emit video/telemetry
  • 🧑‍🦰 Patient: requires treatment

Inputs / Outputs / Ops

  • Inputs: 📞 alarm · 🗺️ map · 🚗 traffic
  • Outputs: 📝 tasks · 🛣️ route · ⏰ SLA
  • Ops: spawn → route → emit
📋 Emergency Scenario: Child Fall in Park

Scenario Description: A child has fallen and injured themselves in a busy park. Multiple agents need to coordinate to assess the situation, provide immediate aid, and ensure proper medical response.

T+0s
Emergency Detected: Vision agents detect child fall incident
T+2s
Assessment: Master agent coordinates situation assessment
T+5s
Response: Action agents provide first aid and call emergency services
T+10s
Resolution: Emergency services arrive, child receives proper care
🔧 DSL Primitives Used in Emergency Response

1. Task Creation and Assignment

ATSLP (Adaptive Task Scheduling with Load Prediction)
// Create emergency assessment task Task emergency_assessment = TaskBuilder() .name("assess_child_injury") .prompt("Assess the severity of child's injury and determine immediate actions needed") .agent("vision_agent_1") .priority(CRITICAL) .timeout(5.0) .build();
HCMPL (Hierarchical Cache Management with Pattern Learning)
// Cache emergency response patterns CacheManager.cache_pattern("child_fall_response", { "immediate_actions": ["assess_injury", "call_emergency", "comfort_child"], "medical_priority": "high", "response_time_limit": 10.0 });
CALK (Collaborative Agent Learning with Knowledge Transfer)
// Share emergency response knowledge KnowledgeTransfer.share_experience( from_agent="experienced_medic_agent", to_agent="new_volunteer_agent", knowledge_type="emergency_first_aid", context="child_injury_scenario" );
👁️ Vision Agents
  • Detect emergency situation
  • Assess injury severity
  • Identify location and context
  • Monitor crowd safety
  • Track emergency response progress
🧠 Master Agent
  • Coordinate emergency response
  • Prioritize tasks and resources
  • Manage agent communication
  • Make critical decisions
  • Monitor overall situation
🚑 Action Agents
  • Provide immediate first aid
  • Call emergency services
  • Clear area for medical access
  • Comfort and reassure child
  • Coordinate with medical personnel

🎮 Emergency Scenario Demo Controls

Ready to start emergency scenario
0.0
Response Time (s)
0
Active Agents
0
Tasks Completed
0%
Success Rate
Emergency Response Timeline
Agent Activity During Emergency
Emergency Response System Ready
DSL Framework Initialized with Emergency Primitives
Click "Start Emergency Scenario" to begin simulation
🗺️ Live Map · SLA Monitor · DSL Stepper

Map & Routes

SLA: 00:30 remaining

DSL Program (Stepper)

 spawn vision_agent id=V1 caps=["detect","track"]
 spawn master_agent id=M caps=["plan","allocate"]
 spawn action_agent id=A1 caps=["evacuate","medic"]
 with_sla { deadline:30s, retry:1 }
   route task assess(location=Playground) to V1
   route task plan_evac(zone=NorthGate) to M
   gather { route task carry_kit() to A1 } reduce=concat
 end
 emit incident_detected { loc:Playground, severity:HIGH }
step 0/0

Raw Data (run.json)

📊 Real Performance Data from Emergency Simulation

Based on actual testing with the park emergency scenario: