Information Technology Reference
In-Depth Information
can be obtained. Specifically, we found it most effective when this number just covers
the quantity the agent needs in low demand games. Thus on Day 0, SouthamptonSCM
orders a large number of components (2000, 2000, 2500, 3500, 5000) from each sup-
plier with corresponding delivery dates of Day 10, 25, 40, 70 and 110. These dates were
chosen in order to give the agent a steady stream of components for the early to middle
part of the game. The agent accepts the corresponding offer if the delivery date is not
too far from the date it asks for. However, if the demand turns out to be greater than
what the agent ordered, it can still buy components (at higher prices) during the rest of
the game. In particular, after the Day 0 order, the agent keeps asking for small quanti-
ties of components from the suppliers and placing orders for them if the offer price is
low. At about Day 140, the agent starts to order components for the rest of the game.
It does this based on the average daily demand for computers (as a predictor of how
many components are needed) and buys gradually if the offer prices from the suppliers
are low.
3.2
The Customer Agent
The customer agent is the key component in SouthamptonSCM's strategy (because we
believe that offering the appropriate price at the right time is vital for success). If the
price is too low, the agent will receive a low profit and if it is too high it will fail to win
any orders (because customers always choose the lowest offer price among those they
receive). Here, the key challenges are to determine which customer RFQs to bid for and
at what price. To achieve this, we use inventory driven methods to choose RFQs and
soft computing techniques to calculate the price (see below).
Choosing RFQs and setting prices.
The customer agent uses an inventory driven
strategy when selecting customer RFQs. That is, it only offers customers PCs according
to what is presently available in its inventory. By doing this, the agent avoids getting
penalties for committing to more than it can produce (the quantity of PCs it can produce
is constrained by the availability of components and factory cycles).
In more detail, Table 1 shows the strategy we use. Given a customer RFQ (
i
,
q
,
p
res
,
c
penalty
,
d
due
),where
i
is the type of PC the customer wants,
q>
0 the
quantity,
p
res
>
0 the reservation price (maximum it will pay),
c
penalty
>
0 the fine
if the computers are not delivered on time, and
d
due
the desired delivery date. On each
day, the customer agent receives a bundle of such RFQs and sorts them in the order
of decreasing (
p
res
−
∈{
1
,
···
,
16
}
c
penalty
/q
). The intuition here is that the agent will first serve
customers with high reserve prices and low penalties. This is because the higher the
p
res
, the more profit will be made (compared to selling the same product to a customer
with a low
p
res
). At the same time, the agent also wants to avoid getting high penalty
orders because of the inherent uncertainties that exist in the game.
The next consideration relates to the agent's production capacity. Specifically, as
there is only limited production capacity per day, the agent needs to calculate the num-
ber of cycles that can be offered to respond to the customer RFQs of that day.
2
Thus, it
2
Note here the agent does not offer the exact number of cycles that are available (
C
[
d
due
−
2])
on day (
d
due
−
2), but rather it includes a risk factor (
λ × C
[
d
due
−
2]) which enables it to
offer more than it actually has in order to maximise the production utilisation. Here
λ>
1.
