Mitigating the risks from small UAVs

Earlier this week, the FAA released proposed rules for operators of small UAVs (drones).  Small UAVs are those that weigh less than 55 pounds. 

From a risk management perspective, the rules propose a variety of preventive actions: fly only in daylight, do not fly over people, do not fly in bad weather, fly below 500 feet altitude, fly at speeds less than 100 miles per hour, and do not fly in airport flight paths and restricted airspace areas.  Most of the rules are meant to prevent accidents by reducing the likelihood of losing control, colliding with other aircraft, and crashing into third persons on the ground. 

The rules do not propose any contingency plans to minimize risk when the operator loses control of the UAV.  Indeed, perhaps the only ones that can be considered are fanciful ones like a "disassemble" command that makes the UAV divide into smaller pieces that would cause less damage on impact or deploying airbags like those on the Mars Pathfinder.  A more reasonable contingency might be a siren and flashing light that are activated to warn those nearby when the UAV begins to crash (like shouting "fore" when a golfer drives a ball towards unsuspecting bystanders).

Compensation and choice strategies for screening

In multicriteria decision situations, many choice strategies are used for selecting one alternative from the set of available alternatives, which is the essence of decision making.  In some cases, however, the decision-maker first wants to screen the available alternatives to find a subset to consider in more detail. Choice strategies can be used for screening as well as for selection.  In particular, the satisficing and disjunctive choice strategies are suited for screening.  The satisficing strategy sets a threshold or cutoff for each attribute and keeps only alternatives that satsify all of those thresholds.  The disjunctive strategy also sets a threshold for each attribute, but it  keeps any alternative that satisfies at least one of those thresholds.

These strategies correspond to preferences for non-compensating and compensating solutions.  A decision-maker who prefers non-compensating solutions will want solutions that are good in every way, which the satisficing strategy identifies.  A decision-maker who prefers compensating solutions may set higher thresholds and use a disjunctive strategy to screen the alternatives. Although no alternative can satisfy all of the thresholds, the decision-maker will be happy with those alternatives that can satisfy at least one of them because they are compensating solutions: each one performs well on at least one attribute, which compensates for poor performance on the other attributes. 


For example, suppose Joe and Rose are considering which college each one should attend.  Joe prefers non-compensating solutions and uses a satisficing strategy: he wants a college that has a reasonable tuition, that is at most 200 miles from home, AND has a top-40 engineering program.  He will narrow his search to colleges that meet all of those criteria.  Rose, however, prefers compensating solutions and uses a disjunctive strategy: she wants a college that is very inexpensive OR within 75 miles of home OR has a top-10 engineering program.  Her criteria are more difficult to meet, but she will narrow her search to those colleges that can meet any one of her criteria, which will be a very different set than those that Joe considers.  For instance, Joe would not consider a top-10 school that is far away and expensive, but Rose would.