Evaluating Predictive Models in the Orienteering Problem with Stochastic Profits: A Simulation Study

Purpose - The aim of this paper is to test the effectiveness statistical model selection measures in terms decision quality for orienteering problem with stochastic profits using simulation. Design/methodology/approach – This based on a quantitative numerical approach where various are evaluated computational experiments including modelbased computer-generated random data. Findings findings include experimental results showing deficiency about 6.5 units classical relative decision-based measure Tsiligirides benchmark instances. Discussion While suitable accuracy reasons, misspecified models sometimes do lead better outcomes. From practical perspective, order carry out prescriptive analytics problems, having access reasonable algorithm at prediction stage data-analysis can be beneficial downstream realized profit.