Thermal agents: an application of genetic programming to virtual engineering

The temperature profile across an object is easy to compute by iterative methods. The time spent waiting for iterative solutions to converge for multiple objects in a complex configuration is an impediment to exploratory analysis of engineering systems. A rapidly computed initial guess can speed convergence for an iterative thermal solver. We describe and test a system for creating thermal agents that supply such initial guesses. Thermal agents are specific to an object geometry but general across different thermal boundary conditions. During an offline training phase, genetic programming is used to locate a thermal agent by training on one or more sets of boundary conditions. In use, thermal agents transform boundary conditions into a rapidly converged set of initial values on a cellular decomposition of an object.