Information Technology Reference
In-Depth Information
before, we attribute the success of our agent to the adaptive control of the offering price
and this is what we are most interested in here. Thus, we decided to analyse how the
pricing works compared with other methods. To do this, we devised two competitor
agents that adopt identical strategies to SouthamptonSCM except for the method they
use to offer prices. The alternative methods
8
we consider are consistent with the broad
classes of behaviour that were adopted by several of the agents in the competition:
- High-Price agent (HP-agent).
This agent bids aggressively at high offer prices to
obtain a higher profit margin in selling the PCs. It will take the risk of stocking
a large number of PCs and components in the factory and paying storage cost for
them. But when its PCs are sold they fetch high prices and mean it can very quickly
build up profits. In more detail, the prices that HP-agent offer are the
maximum
of
the cost of the computer plus a fixed profit margin (here it is 300) and the comput-
ers' reserve price minus 1. Thus, when the reserve price is high enough, it sells at
reserve price minus 1. Otherwise, it sells at the price of the cost of components plus
a fixed margin.
- High-Volume agent (HV-agent).
This agent bids cautiously and only seeks to at-
tain a reasonable profit margin for each order, but it tries to maximise the volume of
orders. This means that the agent wants to sell its PCs quickly and it does not want
to take the risk of stocking components or PCs (especially in games with low cus-
tomer demand). Specifically, it offers the computers at the cost of the computer plus
a small margin. Here the margin is set to 300 in the first 180 days of the game and
this is then decreased to 0 linearly till the end of the game. This policy is adopted
because the agent hopes that it can sell all the computers by the end of the game.
Besides these two kinds of agents, the other competing participants are the dummy
agents provided by the organisers. These use a build-to-order method and offer prices
which are chosen uniformly from 80
100% of the reserve prices. Generally, the
dummy agent can be viewed as being high-volume because it often offers a low price
(but it differs from our HV-agent in that it uses the build-to-order method). Given this
background, three groups of experiments were conducted to examine the performance
of each kind of agent in various situations. In experiment A, there is one Southamp-
tonSCM, one HP-agent, one HV-agent and three dummy agents. In experiment B, we
increase the number of HP-agents to 2 and decrease the number of dummy agents to 2.
In experiment C the number of HP-agents is 3 and the number of dummies is 1. The
average revenue of each kind of agent in each of the experiments are then plotted.
9
−
8
Although neither alternative we chose employs market history information (
i.e.
the highest
and lowest transaction prices of the previous day), they represent the two common attitudes
for setting bidding prices: “High-Price” and “High-Volume”.
9
Note that the dummy agents in the game represent default players. We aim to evaluate the
performance of the other three types of agents instead of dummy agents. In order to obtain
valid results that represent realistic games, we varied the number of dummy agents and in
some games, there is only one dummy agent (e.g. experiment C). We believe this environment
is able to represent realistic games since even in real games there are some players who use
build-to-order like strategies.
