https://en.wikipedia.org/wiki/Agent-bas ... _economics
For context, and perhaps for interest: it is an subsistence economy in that primary producers grow things (crops/food) and sell them to consumers that need those crops to eat and survive. The driver for each agent is to earn money so they can buy food and survive. There is a 'tier 2' where primary producers can buy luxury items not essential to survival. This is represented by buying wood from wood-choppers. The wood is not essential for survival, but is a step towards the building of a house. My simulator allows for agents to provide a service - carpenters. They will take the wood purchased by another agent and build a house for that agent. The house is not essential for survival but living in a house is more comfortable and provides efficiency and productivity, and therefore better places that agent to be fiscally viable. Additionally, the simulator charges a goods and services tax (GST) on every transaction. An agent can be a tax-collector that collects this GST. The GST goes to the 'government' and the government can hire a public-servant to turn the GST 'income' into welfare payments. The welfare payments go to the unemployed and those that can't afford to buy food necessary for survival - thereby - in a sense, completing the circle.
In it's purest form, an ACE is a 'set-and-watch' simulator where you set initial values for research and experimentation and watch the sim play out. I'm gamefying my ACE so that it is ... well ... a game!
In summary, the simulation currently supports: primary producers, wood-choppers, carpenters, tax collectors, welfare officers. Assets/commodities include farms, wood, houses, GST, welfare.
Agents trade commodities between themselves to survive.
And finally - to my question:
With all the above "engine" already built and developed, I'd like to develop a market/economy with the following principles:
- agents can determine needs and wants and can prioritise in a way that ensures their survival (this is already simulated)
- agents transact with other agents directly (no central 'clearing house') (this is already simulated)
- agents have an imperfect view of the world but they do have a memory of previous trades/transactions (this is already simulated)
- agents that are not 'fiscally viable' (starve) will exit the simulation, taking assets and knowledge with them (already simulated)
- agents will randomly join the simulation with minimal assets and default knowledge (already simulated)
- agents that are buying (making a 'bid') need to somehow determine the most they will pay (need help with this)
- agents that are selling (making an 'ask') need to somehow determine the lowest they will offer (need help with this)
I'm essentially trying to work out pricing based on supply and demand (I think) so the agents have some chance of transacting in a way that is not detrimental to either agent and therefore not compromising/destroying the economy/simulation.
Lots of words. I'm not an economics student but I have spent the last few weeks reading countless research papers and I thought I'd see if I can inspire some brain-cells in this community.
Any thoughts and guidance appreciated. Thanks.