Information Technology Reference
In-Depth Information
Day vs. ASP
Market Demand vs. ASP
1.4
1.4
1.2
1.2
1
1
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0
0
0
50
100
150
200
250
0
50
100
150
200
250
300
350
400
Simulation Day
Market Demand
(a)
(b)
Bid Density vs. ASP
Manufacturer PC Inventory vs. ASP
1.4
1.4
1.2
1.2
1
1
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0
0
0
1
2
3
4
5
6
7
0
500
1000
1500
2000
2500
3000
3500
4000
Bid Density
Manufacturer PC Inventory
(c)
(d)
Fig. 5.
Plots showing the relationships between four variables and market ASPs in SCM-04.
10000 randomly selected points are shown for each variable (out of approximately 45000 to-
tal). The dark lines on each plot are the average ASPs at each value for discrete variables or
within small bins for continuous variables.
sufficient to explain 65% of the variance in ASPs. This surprisingly powerful result that
speaks to the power of the basic forces of supply and demand in the SCM markets.
We take this analysis a step further by considering the features of the markets that
individual agents bid and sold PCs in. Table 6 gives average values for each of these four
factors for each of the PCs sold and PCs bid on by each agent. The two agents with the
highest ASPs were
Deep Maize
and
FreeAgent
. These agents sold PCs earlier in the
game, in markets with the lowest PC inventory, highest demand, and low bid densities;
all of these are associated with high market ASPs.
UMTac-04
had the lowest ASP by
a wide margin. This agent sold later than most agents in markets with low demand and
high PC inventory levels; all of these are associated with low market ASPs. The other
agents have mid-range ASPs, and seem to have a mix of factors working for and against
them. For instance,
SouthamptonSCM
sells exceptionally early, but in lower demand
markets with relatively high inventory levels.
5
Conclusions: Agent Performance
We have presented and analyzed data on many aspects of the TAC-04/SCM final round,
which has revealed interesting details about the different strategies employed by the
agents. We can now offer a reasonably compelling explanation for why the agents fin-
