Abstract # 200:

Scheduled for Monday, September 15, 2014 02:15 PM-02:30 PM: Session 27 (Mary Gay) Oral Presentation


WILD VERVETS ARE “SELLING SNAKE OIL” BY SOLVING ROUTING PROBLEMS WITH SIMPLE HEURISTICS

J. A. Teichroeb
Dept. of Evolutionary Anthropology, Duke University, Rm 103B Biological Sciences Building, Box 90383 , Durham, NC 27708, USA
line
     The “Traveling Salesman Problem” (TSP) requires individuals to visit several targets using the shortest possible path and becomes mathematically impossible with many destinations. Most non-human primate species have shown limited ability to route plan. However, captive vervets were shown to solve a TSP for six sites, while appearing to plan three steps ahead (Cramer & Gallistel, 1997). I investigated the abilities of wild vervets in Uganda to solve an open-TSP routing problem for six, equally-rewarding experimental platforms. The arrangement of platforms allowed it to be determined whether vervets found the shortest route or used one of three simple heuristics (nearest-neighbor rule, additive gravity, or the convex-hull) to navigate. Single vervet’s paths were consistent with simple heuristics 61.9% (n=276 trials, 9 ind.) of the time and inefficient routes 42.8% of the time. The convex-hull heuristic was most efficient, leading to the shortest possible paths, and routes were consistent with this heuristic most often (38.7%). When two or more foragers competed in the experiment (n=59 trials, 13 ind.), paths consistent with the most efficient heuristics in a competitive situation (nearest-neighbor rule and additive gravity) increased and use of the convex-hull decreased (P=0.002). Thus, vervets often appear to solve the TSP because they use the most efficient simple heuristic, leading to an optimal route. Hence vervets are not actually traveling salesmen but are rather “selling snake oil”.