David Andre

Automated synthesis of analog electrical circuits by means of genetic programming

Analog circuit synthesis entails the creation of both the topology and the sizing (numerical values) of all of the circuits components. This paper presents a single uniform approach using genetic programming for the automatic synthesis of both the topology and sizing of a suite of eight different prototypical analog circuits, including a low-pass filter, a crossover filter, a source identification circuit, an amplifier, a computational circuit, a time-optimal controller circuit, a temperature-sensing circuit, and a voltage reference circuit. The problem-specific information required for each of the eight problems is minimal and consists of the number of inputs and outputs of the desired circuit, the types of available components, and a fitness measure that restates the high-level statement of the circuits desired behavior as a measurable mathematical quantity. The eight genetically evolved circuits constitute an instance of an evolutionary computation technique producing results on a task that is usually thought of as requiring human intelligence.

Evolving Team Darwin United

The RoboCup simulator competition is one of the most challenging international proving grounds for contemporary AI research. Exactly because of the high level of complexity and a lack of reliable strategic guidelines, the pervasive attitude has been that the problem can most successfully be attacked by human expertise, possibly assisted by some level of machine learning. This led, in RoboCup97, to a field of simulator teams all of whose level and style of play were heavily influenced by the human designers of those teams. In contrast, our 1998 team was designed entirely by the process of genetic programming. Our evolved team placed in the middle of the pack at Robocup98, despite the fact that it was largely machine learned rather than hand coded. This paper presents our motivation, our approach, and the specific construction of our team that created itself from scratch.

Automated Design of Both the Topology and Sizing of Analog Electrical Circuits Using Genetic Programming

This paper describes an automated process for designing analog electrical circuits based on the principles of natural selection, sexual recombination, and developmental biology. The design process starts with the random creation of a large population of program trees composed of circuit-constructing functions. Each program tree specifies the steps by which a fully developed circuit is to be progressively developed from a common embryonic circuit appropriate for the type of circuit that the user wishes to design. The fitness measure is a user-written computer program that may incorporate any calculable characteristic or combination of characteristics of the circuit. The population of program trees is genetically bred over a series of many generations using genetic programming. Genetic programming is driven by a fitness measure and employs genetic operations such as Darwinian reproduction, sexual recombination (crossover), and occasional mutation to create offspring. This automated evolutionary process produces both the topology of the circuit and the numerical values for each component. This paper describes how genetic programming can evolve the circuit for a difficult-to-design low-pass filter.

A parallel implementation of genetic programming that achieves super-linear performance

This paper describes the successful parallel implementation of genetic programming on a network of processing nodes using the transputer architecture. With this approach, researchers of genetic algorithms and genetic programming can acquire computing power that is intermediate between the power of currently available workstations and that of supercomputers at intermediate cost. This approach is illustrated by a comparison of the computational effort required to solve a benchmark problem. Because of the decoupled character of genetic programming, our approach achieved a nearly linear speed up from parallelization. In addition, for the best choice of parameters tested, the use of subpopulations delivered a super-linear speed-up in terms of the ability of the algorithm to solve the problem. Several examples are also presented where the parallel genetic programming system evolved solutions that are competitive with human performance.

Automated synthesis of computational circuits using genetic programming

Analog electrical circuits that perform mathematical functions (e.g., cube root, square) are called computational circuits. Computational circuits are of special practical importance when the small number of required mathematical functions does not warrant converting an analog signal into a digital signal, performing the mathematical function in the digital domain, and then converting the result back to the analog domain. The design of computational circuits is difficult even for mundane mathematical functions and often relies on the clever exploitation of some aspect of the underlying device physics of the components. Moreover, implementation of each different mathematical function typically requires an entirely different clever insight. This paper demonstrates that computational circuits can be designed without such problem-specific insights using a single uniform approach involving genetic programming. Both the circuit topology and the sizing of all circuit components are created by genetic programming. This uniform approach to the automated synthesis of computational circuits is illustrated by evolving circuits that perform the cube root function (for which no circuit was found in the published literature) as well as for the square root, square, and cube functions.

Mobile robot obstacle avoidance via depth from focus

A critical challenge in the creation of autonomous mobile robots is the reliable detection of moving and static obstacles. In this paper, we present a passive vision system that recovers coarse depth information reliably and efficiently. This system is based on the concept of depth from focus, and robustly locates static and moving obstacles as well as stairs and dropoffs with adequate accuracy for obstacle avoidance. We describe an implementation of this vision system on a mobile robot as well as real-world experiments both indoors and outdoors. These experiments have involved several hours of continuous and fully autonomous operation in crowded, natural settings.

Reuse, Parameterized Reuse, and Hierarchical Reuse of Substructures in Evolving Electrical Circuits Using Genetic Programming

Most practical electrical circuits contain modular substructures that are repeatedly used to create the overall circuit. Genetic programming with automatically defined functions and the recently developed architecture-altering operations provides a way to build complex structures with reused substructures. In this paper, we successfully evolved a design for a two-band crossover (woofer and tweeter) filter with a crossover frequency of 2,512 Hz. Both the topology and the sizing (numerical values) for each component of the circuit were evolved during the run. The evolved circuit contained three different noteworthy substructures. One substructure was invoked five times thereby illustrating reuse. A second substructure was invoked with different numerical arguments. This second substructure illustrates parameterized reuse because different numerical values were assigned to the components in the different instantiations of the substructure. A third substructure was invoked as part of a hierarchy, thereby illustrating hierarchical reuse.

Evolution of a 60 Decibel Op Amp Using Genetic Programming

Genetic programming was used to evolve both the topology and sizing (numerical values) for each component of a low-distortion, low-bias 60 decibel (1000-to-1) amplifier with good frequency generalization.

Synthesis of topology and sizing of analog electrical circuits by means of genetic programming

Abstract The design (synthesis) of an analog electrical circuit entails the creation of both the topology and sizing (numerical values) of all of the circuits components. There has previously been no general automated technique for automatically creating the design for an analog electrical circuit from a high-level statement of the circuits desired behavior. This paper shows how genetic programming can be used to automate the design of eight prototypical analog circuits, including a lowpass filter, a highpass filter, a bandstop filter, a tri-state frequency discriminator circuit, a frequency-measuring circuit, a 60 dB amplifier, a computational circuit for the square root function, and a time-optimal robot controller circuit.

Evolving computer programs using rapidly reconfigurable field-programmable gate arrays and genetic programming

This paper describes how the massive parallelism of the rapidly reconfigurable Xilinx XC6216 FPGA (in conjunction with Virtual Computings H.O.T. Works board) can be exploited to accelerate the time-consuming fitness measurement task of genetic algorithms and genetic programming. This acceleration is accomplished by embodying each individual of the evolving population into hardware in order to perform the fitness measurement task. A 16-step sorting network for seven items was evolved that has two fewer steps than the sorting network described in the 1962 OConnor and Nelson patent on sorting networks (and the same number of steps as a 7-sorter that was devised by Floyd and Knuth subsequent to the patent and that is now known to be minimal). Other minimal sorters have been evolved.