Real-time AI Research Articles

A Survey of Real-Time Strategy Game AI Research and Competition in StarCraft

This paper presents an overview of the existing work on AI for real-time strategy (RTS) games. Specifically, we focus on the work around the game StarCraft, which has emerged in the past few years as the unified test bed for this research. We describe the specific AI challenges posed by RTS games, and overview the solutions that have been explored to address them. Additionally, we also present a summary of the results of the recent StarCraft AI competitions, describing the architectures used by the participants. Finally, we conclude with a discussion emphasizing which problems in the context of RTS game AI have been solved, and which remain open.

Flexible hard real-time scheduling for deliberative AI systems

Constructing deliberative real-time AI systems is challenging due to the high execution-time variance in AI algorithms and the requirement of worst-case bounds for hard real-time guarantees, often resulting in poor use of system resources. Using a motivating case study, the general problem of resource usage maximization is addressed. We approach the issues by employing a hybrid task model for anytime algorithms, which is supported by recent advances in fixed priority scheduling for imprecise computation. In particular, with a novel scheduling scheme based on Dual Priority Scheduling, hard tasks are guaranteed by schedulability analysis and scheduled in favor of optional and anytime components which are executed whenever possible for enhancing system utility. Simulation studies show satisfactory performance on the case study with the application of the scheduling scheme. We also suggest how aperiodic tasks can be scheduled effectively within the framework and how tasks can be prioritized based on their utilities by an efficient algorithm. These works form a comprehensive package of scheduling model, analysis, and algorithms based on fixed priority scheduling, providing a versatile platform where real-time AI applications can be suitably facilitated.

Synchronization of ovulation in cyclic gilts with porcine luteinizing hormone (pLH) and its effects on reproductive function

The overall objective was to evaluate the use of porcine luteinizing hormone (pLH) for synchronization of ovulation in cyclic gilts and its effect on reproductive function. In an initial study, four littermate pairs of cyclic gilts were given altrenogest (15 mg/d for 14 d). Gilts received 500 μg cloprostenol (Day 15), 600 IU equine chorionic gonadotropin (eCG) (Day 16) and either 5 mg pLH or saline (Control) 80 h after eCG. Blood samples were collected every 4 h, from 8 h before pLH/saline treatment to the end of estrus. Following estrus detection, transcutaneous real-time ultrasonography and AI, all gilts were slaughtered 6 d after the estimated time of ovulation. Peak plasma pLH concentrations (during the LH surge), as well as the amplitude of the LH surge, were greater in pLH-treated gilts than in the control ( P = 0.01). However, there were no significant differences between treatments in the timing and duration of estrus, or the timing of ovulation within the estrous period. In a second study, 45 cyclic gilts received altrenogest for 14–18 d, 600 IU eCG (24 h after last altrenogest), and 5 mg pLH, 750 IU human chorionic gonadotropin (hCG), or saline, 80 h after eCG. For gilts given pLH or hCG, the diameter of the largest follicle before the onset of ovulation (mean ± S.E.M.; 8.1 ± 0.2 and 8.1 ± 0.2 mm, respectively) was smaller than in control gilts (8.6 ± 0.2 mm, P = 0.05). The pLH and hCG groups ovulated sooner after treatment compared to the saline-treated group (43.2 ± 2.5, 47.6 ± 2.5 and 59.5 ± 2.5 h, respectively; P < 0.01), with the most synchronous ovulation ( P < 0.01) in pLH-treated gilts. Embryo quality (total cell counts and embryo diameter) was not significantly different among groups. In conclusion, pLH reliably synchronized ovulation in cyclic gilts without significantly affecting embryo quality.

Preliminary Thoughts on the Application of Real-Time AI Game-Tree Search to Control

We introduce a formal definition of a general hybrid system control game, and a decision procedure for an agent which uses simulation to inform game-playing decisions in real-time. We discuss heuristics informing the estimation of expected utilities of actions, describe metalevel reasoning criteria which direct game-tree expansion according to decision relevance, and point out challenges in applying ideas from AI real-time game-tree search to control.

A Rule-Based Real-Time AI Problem Solving Mechanism

A Rule-Based Real-Time AI Problem Solving Mechanism

Asynchronous superimposition mechanisms of concurrent competitve waves for hyper-distributed hyper-parallel heuristic problem solving

This paper presents a new approach to hyper-distributed hyper-parallel heuristic AI problem solving, which is based on asynchronous superimposition of synchronous homogeneous concurrent propagations of competitive waves. In comparison with synchronous homogeneous mechanism, the proposed approach shows better generality, suitability and feasibility for real-time AI processing, especially for the search of implicit AND/OR graphs.

World modeling for the dynamic construction of real-time control plans

As intelligent, autonomous systems are embedded in critical real-world environments, it becomes increasingly important to rigorously characterize how these systems will perform. Research in real-time computing and control has developed ways of proving that a given control system will meet the demands of an environment, but has not addressed the dynamic planning of control actions. Building an agent that can flexibly achieve its goals in changing environments requires a blending of real-time computing and AI technologies. The Cooperative Intelligent Real-time Control Architecture (CIRCA) implements this blending by executing complex AI methods and guaranteed real-time control plans on separate subsystems. We describe the formal model of agent/environment interactions that CIRCA uses to build control plans, and we show how those control plans are guaranteed to meet domain requirements. CIRCA's world model provides the information required to make real-time performance guarantees, but avoids unnecessary complexity.

The challenges of real-time AI

The research agendas of artificial intelligence and real-time systems are converging as AI methods move toward domains that require real-time responses, and real-time systems move toward complex applications that require intelligent behavior. They meet at the crossroads in an exciting new subfield commonly called "real-time AI." This subfield is still being defined, and the precise goals for various real-time AI systems are in flux. Our goal is to identify promising areas for future research in both real-time and AI techniques. We describe an organizing conceptual structure for current real-time AI research, exploring the meanings this term has acquired. We then identify the goals of real-time AI research and specify some necessary steps for reaching them.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Trends in artificial intelligence applications for real-time control (a speculative study)

Some of the theory-bound evolutional trends in real-time AI applications are pointed out based on analysis of essential properties of real-time systems as well as AI based systems. The evolutional mainstream is increasing interdisciplinary integration. Three subtrends are illustrated on examples: mechanical combination of methods, AI methods used for approximate solution of “classical” problems, and abstract methods applied in new domains. In addition similarity between integrated circuits and real-time systems design and increased use of formal verification at the early stages of systems development are pointed out.

Continuation compilation for concurrent logic programming

A new and powerful approach to threading is proposed, that is designed to improve the responsiveness of concurrent logic programs for distributed, real-time AI applications. The technique builds on previously proposed scheduling techniques to improve responsiveness by synchronously passing control and data directly from a producer to a consumer. Furthermore, synchronous transfer of data requires less buffering and so less garbage is produced. Arguments are also passed in registers, further reducing overheads.

Integrating real-time AI techniques in adaptive intelligent agents

Adaptive intelligent agents pursue multiple goals in dynamic, complex, uncertain environments. Illustrative applications include medical monitoring agents, autonomous robots, and animated actors. Real-time considerations include hard and soft deadlines, as well as more qualitative constraints. There are no bounded-time algorithms for such applications. Agents must construct goal-seeking behavior opportunistically out of component behaviors triggered at runtime. In contrast to conventional real-time systems, an agent's real-time effectiveness cannot be guaranteed by a single provably optimal technique. Instead, it must emerge from interactions among multiple heuristic techniques at architectur, cognition, and control levels.

191 Top-down design of embedded real-time AI systems

191 Top-down design of embedded real-time AI systems

A REAL-TIME CONTROL ARCHITECTURE FOR AN APPROXIMATE PROCESSING BLACKBOARD SYSTEM

Approximate processing is an approach to real-time AI problem solving in domains in which compromise is possible between the resources required to generate a solution and the quality of that solution. It is a satisficing approach in which the goal is to produce acceptable solutions within the available time and computational resource constraints. Previous work has shown how to integrate approximate processing knowledge sources within the blackboard architecture. However, in order to solve real-time problems with hard deadlines using a blackboard system, we need to have: (1) a predictable blackboard execution loop, (2) a representation of the set of current and future tasks and their estimated durations, and (3) a model of how to modify those tasks when their deadlines are projected to be missed, and how the modifications will affect the task durations and results. This paper describes four components for achieving this goal in an approximate processing blackboard system. A parameterized low-level control loop allows predictable knowledge source execution, multiple execution channels allow dynamic control over the computation involved in each task, a meta-controller allows a representation of the set of current and future tasks and their estimated durations and results, and a real-time blackboard scheduler monitors and modifies tasks during execution so that deadlines are met. An example is given that illustrates how these components work together to construct a satisficing solution to a time-constrained problem in the Distributed Vehicle Monitoring Testbed (DVMT).

Dice: A Real-Time Toolbox

In this paper research the DICE toolbox is described. This toolbox has been developed for real-time AI issues. DICE offers progressive reasoning as a way to handle time critical situations. DICE is a multitasking environment developed for a VAX/VMS environment. It offers both boolean and multivalued logic implementations. Some examples concerning the syntax are given.

Top-down design of embedded real-time AI systems

A formal method is proposed for the specification and verification of embedded real-time systems. We consider distributed systems in which parallel processes communicate by sending messages along synchronous or asynchronous channels. To verify that a program satisfies a specification, a compositional proof system is given. Compositionality enables verification during the process of top-down program design. This is illustrated by an example of a rail road crossing. Starting from an assertional specification, we design a program that controls lamps and barriers using information from sensors that signal the passage of trains at a certain distance from the crossing.

Dice: A real-time toolbox

In this paper research the DICE toolbox is described. This toolbox has been developed for real-time AI issues. DICE offers progressive reasoning as a way to handle time critical situations. DICE is a multitasking environment developed for a VAX/VMS environment. It offers both boolean and multivalued logic implementations. Some examples concerning the syntax are given.

Real-time AI

Real-time AI

Artificial intelligence in hard real-time — a new paradigm needed?

This essay is to point out some of the basic differences between conventional and real-time applications of AI and to provoke discussion on whether or not a new paradigm would facilitate better understanding of these differences. Some of the listed differences are illustrated by examples from references. An attempt has been made to project existing AI methods onto the deduction-induction axis so as to find an analogy with human behaviour under hard real-time constraints, e.g. reflexes. The conclusion is that the developers of real-time AI methods might benefit by using a “reflex based” paradigm.

Real-time continuous AI systems

Many applications of artificial intelligence to practical problems in industry require systems that run continuously and in real time. The paper describes our experiences from developing a number of such systems in the domains of air defence, ship positioning and online monitoring and diagnosis. The problems that arose in these systems were due to the inherent dynamism, time and space limitations on the processing possible, uncertainty over the current state of the world, resource limitations and several others. The characteristics of these problems are described in the paper along with their implications for choosing suitable AI techniques. A common theme in the applications we have studied is the loop of plan generation, monitoring, diagnosis and replanning; examples of how these processes can be performed are given and some of the possible pitfalls indicated. No claims are made for a universal panacea for the problems in these difficult domains, but techniques which have proved useful are given. The paper does not aim to be a complete dissertation on the subject of real-time AI systems but should provide a usable introduction.