MINIMAX PREDICTION AND AN EVASION GAME.

Abstract

The problem of determining how an evading target should maneuver to survive a lethal weapon, and reciprocally, how the attacker or marksman should aim and when he should fire is, in its general form, much too complex for a satisfactory solution. In this paper, an idealized version of the problem is formulated and solved as a 2 person zero sum game. The game is shown to have a saddle point with pure strategies. The evader moves in one dimension and is constrained to choose from a class of Gaussian processes. The marksman uses linear prediction theory, observing the complete past history of the evader's motion. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jun 06, 1963
Accession Number
AD0415434

Entities

People

  • J. Bram

Organizations

  • Center for Naval Analyses

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Computing-Related Activities
  • Data Science
  • Gaussian Processes
  • Information Science
  • Interdisciplinary Science
  • Maneuvers
  • Mathematical Analysis
  • Mathematics
  • Military Operations
  • Probability
  • Statistical Analysis

Readers

  • Game Theory.
  • Neural Network Machine Learning.
  • Theoretical Analysis.