engleski

Improvement of Political Campaigning by a Computer Simulation Model Based on Game Theory

There are numerous areas where the decision-making process can be significantly improved by using computer simulation. This paper presents a simulation model intended for resolving the issue of optimal advertising strategy in political confrontations. It is a well-known fact that voters' opinions are greatly influenced by the media. With this in mind, in any election campaign it is of great importance to properly assess the effects of particular types of promotion, taking into account the opponents' activities and possible reactions. There is important support for solving such problems, namely models developed in the framework of game theory. The model presented here, theoretically and on a hypothetical example, is based on an assumption that two opposing sides are taking part in an election race. Another assumption is that the percentage of votes gained by one contestant is equivalent to the percentage of votes lost by the other contestant. The values of percentage changes are defined as random variables that follow a normal distribution. In order to assess its parameters as correctly as possible, an adequate database needs to be modeled and constructed. Simulated values are used as a basis for formulating the model. Solving it by means of one of the available program applications leads one to the optimal advertising strategy, which is the result of the first simulation set. By repeating the procedure a number of solutions are generated, which allows strategy distributions to be formed. The total optimal strategy is reached by calculating the means of each of the determined distributions. Since it incorporates a quantitative model and the application of a computer system required for its efficient solution and analysis, the construction described in this paper represents a decision support system.

political campaigning ; computer simulation model ; game theory ; database ; optimal advertising strategy ; decision support system

nije evidentirano