# Simple example based upon the WWW

# Scenario: You have a web site which offers a range of products and
# you want to tailor the products offered or advertised to the type of
# person that is visiting your site.  You do not want to interfere
# with their browsing, so the only clues to their personality are the
# products they seem to be interested in.

values: reward
discount: 0.95

states:

S1 # person is type 1
S2 # person is type 2
SN # person is somewhere in between the two types
SX # nobody connected

actions:

A1 # advertise type 1 products
A2 # advertise type 2 products
AN # advertise neutral products

observations:

O1 # page selected is type 1
O2 # page selected is type 2
ON # page selected is type 3
OB # person buys a product
OX # person leaves the site

# Start state
start include: SX

######################################################################
# State transitions

# If no one is at the web site, then the next person to visit is
# equally likely to be of a certain type
T: * : SX : S1 0.3333333333333
T: * : SX : S2 0.3333333333333
T: * : SX : SN 0.3333333333333

# S1: If you advertise products of interest then they are more likely 
# to stay than if you advertise uninteresting products.
T: A1 : S1 : S1   0.80
T: A1 : S1 : SX   0.2

T: A2 : S1 : S1   0.5
T: A2 : S1 : SX   0.5

T: AN : S1 : S1   0.66666667
T: AN : S1 : SX   0.33333333

# S2: If you advertise products of interest then they are more likely 
# to stay than if you advertise uninteresting products.
T: A1 : S2 : S2   0.5
T: A1 : S2 : SX   0.5

T: A2 : S2 : S2   0.80
T: A2 : S2 : SX   0.2

T: AN : S2 : S2   0.66666667
T: AN : S2 : SX   0.33333333

# SN: 
T: A1 : SN : SN   0.5
T: A1 : SN : SX   0.5

T: A2 : SN : SN   0.5
T: A2 : SN : SX   0.5

T: AN : SN : SN   0.75
T: AN : SN : SX   0.25


######################################################################
# Observation probabilities

# Deterministic observation of a disconnection
O: * : SX : OX   1.0

# S1: More likely to look at products in their own interest
O: * : S1 : O1   0.65
O: * : S1 : O2   0.1
O: * : S1 : ON   0.2

# S1: They'll buy a product with some probability, so the longer you
# keep them at the site, the more chance there is they will buy.
O: * : S1 : OB   0.05

# S2: More likely to look at products in their own interest
O: * : S2 : O1   0.1
O: * : S2 : O2   0.65
O: * : S2 : ON   0.2

# S1: They'll buy a product with some probability, so the longer you
# keep them at the site, the more chance there is they will buy.
O: * : S2 : OB   0.05

# SN: More likely to look at products in their own interest
O: * : SN : O1   0.2
O: * : SN : O2   0.2
O: * : SN : ON   0.55

# SN: They'll buy a product with some probability, so the longer you
# keep them at the site, the more chance there is they will buy.
O: * : SN : OB   0.05

######################################################################
# Rewards

R: * : * : * : OB  1.0