If the advantage of the gift economy is partly due to division of labour, (See GiftAndSelfishForaging) can a division of labour, through exchange work as well? So far we've added three kinds of agents that work with exchange : * BarterAgent, * ProfitSeekingBarterAgent * the first, primitive MoneyAgent For Question3? we ignore profit-seeking barter (which is a minor variant of barter) and compare barter and money with gift and selfish foraging.
Results
The results are here : Selfish Foraging Experiment Parameters Economy of 100 agents of type Agent.Agent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 45.440000 Average of surplus generated by the economy : 2246.040000 Average of total production of the economy during it's run : 4079.040000 Mean of standard deviation of wealth : 22.483158 from a mean of 24.753666 (Smaller sDev the more equal the society) Gift Experiment Parameters Economy of 100 agents of type GiftAgent.GiftAgent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 85.280000 Average of surplus generated by the economy : 1206.580000 Average of total production of the economy during it's run : 6024.580000 Mean of standard deviation of wealth : 8.927779 from a mean of 7.050172 (Smaller sDev the more equal the society) Barter Experiment Parameters Economy of 100 agents of type BarterAgent.BarterAgent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 46.540000 Average of surplus generated by the economy : 2231.260000 Average of total production of the economy during it's run : 4095.460000 Mean of standard deviation of wealth : 21.876617 from a mean of 23.996450 (Smaller sDev the more equal the society) Money Experiment Parameters Economy of 100 agents of type MoneyAgent.MoneyAgent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 46.960000 Average of surplus generated by the economy : 2113.180000 Average of total production of the economy during it's run : 3973.780000 Mean of standard deviation of wealth : 19.926870 from a mean of 22.422749 (Smaller sDev the more equal the society)
In this experiment, neither form of exchange is much improvement over selfish foraging. That's funny, because during development, it seemed as if there was some improvement. Question4? adds a new resource : "love" (who says you can't buy it). This makes some difference. Selfish Foraging Experiment Parameters Economy of 100 agents of type Agent.Agent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 28.940000 Average of surplus generated by the economy : 2126.360000 Average of total production of the economy during it's run : 3835.160000 Mean of standard deviation of wealth : 22.527271 from a mean of 24.503851 (Smaller sDev the more equal the society) Gift Experiment Parameters Economy of 100 agents of type GiftAgent.GiftAgent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 78.240000 Average of surplus generated by the economy : 1810.780000 Average of total production of the economy during it's run : 8519.380000 Mean of standard deviation of wealth : 9.611330 from a mean of 7.651773 (Smaller sDev the more equal the society) Barter Experiment Parameters Economy of 100 agents of type BarterAgent.BarterAgent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 31.980000 Average of surplus generated by the economy : 2275.160000 Average of total production of the economy during it's run : 4270.160000 Mean of standard deviation of wealth : 22.215280 from a mean of 23.740025 (Smaller sDev the more equal the society) Money Experiment Parameters Economy of 100 agents of type MoneyAgent.MoneyAgent In a Random network (av. 22 arcs per node) This economy was run 50 times, each time for 30 steps Summaries Mean no. agents alive : 34.520000 Average of surplus generated by the economy : 2175.140000 Average of total production of the economy during it's run : 4295.540000 Mean of standard deviation of wealth : 19.019840 from a mean of 20.954850 (Smaller sDev the more equal the society)
As noted in GiftAndSelfishForaging, adding more types of resources makes it easier for agents to die (there are more shortages that can kill them) and increases the value of interacting with neighbours. Lets have a look at a summary ... | Type | Agents Alive | Production | sDev | Selfish | 28.9 | 3835 | 22.5 | Gift | 78.2 | 8519 | 9.6 | Barter | 31.9 | 4270 | 22.21 | Money | 34.5 | 4295 | 19 Barter is some improvement over selfishness, Money is better than barter, but Gift is still better by a long margin.
Discussion
A few points to note : * there could be a lot of devil in the details. Look at the algorithms for MoneyAgents and BarterAgents to see if there are any obvious flaws or biases here. And a lot of ways money could be implemented. The simple auctions currently used may not be plausible or efficient. * also, the importance of money may be that it enables trade across larger social distances. Because the agents don't buy stuff other than to fulfil their own needs maybe the virtue of trade is being lost. Suggestions please for a plausible TradeAgent? algorithm.
|