The model uses autonomous agents: buyers and sellers who populate a computer simulation. By using agents, the assumptions concern the actions of buyers and sellers rather than assumptions on the level of the entire market. It comprises agent inputs, outputs and the relationships of these with its internal state. The simulation uses statistics which formulate a time-history of the simulated economy. The model insures repeatability and stability of solutions. However, general properties (convergence/scalability and stability) of agent-based systems are not solved. The model is simplified, presenting an idealized market with a software product traded offered by a number of companies. The price is established in the current period, the profit of the previous period is all used for advertising. Buyers are influenced by: code, market share, advertising effort, willingness to buy new codes. For the open source, there is no advertising and no charge.
The code’s characteristics are: price and brand (the quality is not known). The open source is free of charge (not free to modify and develop). There is a price limit for the user, a market share, an advertising budget, and random effects (which influence the willingness to sample new codes). There are only fixed costs, variable costs are zero. For the open source, there are no costs at all, manufacturing and distribution costs are zero (the only cost is the time used for the download and CD-writing). The open source developers use, share and develop source code, but they represent a small part of the software market.