Page 38 - Zoo Animal Learning and Training

10 1 Learning Theory

Schedules of reinforcement can differ in
Generally, schedules of reinforcement can
VetBooks.ir be continuous or intermittent. On a continu- two ways. Firstly, they can differ based on
ous reinforcement schedule, every emitted
responses or the amount of time passed. In
target behaviour is followed by a reinforcer. whether they come from the number of
An example of this in the natural environment ratio schedules, reinforcement depends on the
would be a baby learning to drink from their number of responses made. Ratio schedules
mother’s teat. Every time they suck on the teat are set to deliver reinforcement following a
they will get milk, which increases their particular number of responses. Interval
chances of sucking on the teat in the future. schedules are set to deliver reinforcement
However, schedules of natural contingencies when one response is made after some
are often intermittent: the behaviour isn’t rein­ amount of time has passed. In fixed schedules,
forced every time it occurs. A classic example the number of responses needed to obtain
of an intermittent schedule in nature is bees’ reinforcement is the same every time. The
foraging behaviour. A single bee will visit sev­ number of responses can be 1 or 1000, but
eral different flowers to find nectar. However, if that number is fixed. In variable schedules,
they visit the same flowers, they may not find the number of responses required for rein­
nectar every time because the flower needs forcement varies around some average.
time to fill. We also see this in zoos/aquariums Let’s go over some examples. A keeper
if you vary whether food is put in environmen­ training an elephant to touch a target delivers
tal enrichment items. The animal might check food on a variable ratio 5 (written as VR 5).
the environmental enrichment item every This means that on average, every fifth
time it is placed into the enclosure, but will response, the elephant will receive food. So
only manipulate the item if it is filled with food. the elephant might receive a piece of sweet
Continuous and intermittent reinforcement potato on the first response, sixth response,
can be broken down into four schedules or second response, eighth response, fifth
reinforcement: fixed ratio, variable ratio, fixed response, and the eighth response and so
interval, and variable interval (Table 1.2). on. If we were to train that same elephant

Table 1.2 Reinforcement schedules.

Reinforcement
schedule Definition Example

Fixed interval Reinforcement is delivered at a Turning out the animals to the yard: every morning at
predictable time interval. 10 am the keeper opens the night enclosure door, but the
animal’s behaviour of checking the door to go outside
isn’t reinforced until they check the door after 10 am.
Variable Response is reinforced after an Animal feedings: the time of feeding an animal may
interval interval of time which varies but vary from day to day, but on average a keeper gives
centres around some average food every 4 hours. Therefore, the animal’s response of
amount of time. checking the bowl will not be reinforced until an
average of 4 hours has passed.
Fixed ratio Response is reinforced only Multiple repetitions: you want the animal you are
after a specified number of training to do multiple repetitions of the same
responses. behaviour. Therefore, you deliver reinforcement after
every 2 correct responses.
Variable ratio Response is reinforced after an Laboratory study: the lever in a Skinner box gives a
average number of responses. pellet on average after 20 pulls. Thus the rat might
receive a pellet after 2 pulls or after 15 pulls, but on
average it is 20 pulls of the lever to receive a pellet.