INTELLIGENTLY ADDING INTELLIGENCE

摘要

All decisions in winter service are based on input information. Decision makers trust this information whilst recognising its limitations. A weather forecast is generated using highly complex algorithms; however, human intervention is often required. Although highly reliant on model outputs, quality weather forecasts require the injection of further intelligence.rnDecision making in winter service is also increasingly reliant on algorithms and the quest for higher levels of refinement comes at a cost. In the current economic climate it may not always be affordable to achieve high accuracy, therefore it is important to determine what can be done at lower cost. The fundamental questions are “How much accuracy do we need for different decision making activities?” and “Where is the intelligence best added?”.rnThis paper considers data use in winter service and explores how the injection of intelligence into uncertain data sets can yield useful, lower cost results. It includes an outline of an experimental, long-range forecasting decision support tool that shows promise for salt stock management, resource planning and other benefits outside of winter service.