Dynamical systems arising from infinite time horizon optimization models (Q4528547)

Economics decisions of intertemporal allocations are described by discrete-time dynamical systems. These dynamical systems used are not defined by arbitary chosen difference equations but are obtained as solutions of maximization problems. This will impose severe restrictions on the properties of the dynamical systems. This paper studies cycles and chaos in autonomous (discrete dynamical systems) economic models. The first model is an economic growth model in which households maximize their utility subject to wealth constraints and firms maximize the discounted sum of profits over an infinite horizon period. This decentralized economy is mathematically equivalent to a welfare maximizing model with a representative agent. The authors formulate this representative agent model and study the dynamics of the optimal solution. They start with a simple aggregated model and extend to a multisector model. They show how cyclical and chaotic dynamics may appear in the multisector model. They also point out that a dynamical optimization program they study is equivalent to a dynamic linear programming problem in which chaotic solutions results. The second model they study is an extension of the first model by allowing direct influences of the production activity among agents. Each agent's activity affects the other agents without going through the market. This creates distortions of the resources allocations and therefore the dynamics of equilibrium paths will be affected. In particular there appear multiple equilibrium paths from the same initial point. This indeterminacy is also studied.