Discrete functions over a continuous domain are approximated by discrete functions with fewer levels. These quantizations are endowed with different types of constraints such as monotonicity and variational constraint...
详细信息
Discrete functions over a continuous domain are approximated by discrete functions with fewer levels. These quantizations are endowed with different types of constraints such as monotonicity and variational constraints. Quantization functions exactly or approximately minimize the squared error to a given discrete function. Exact algorithms are derived from dynamic programming with finite horizon. All algorithms have polynomial run time. (C) 2003 Elsevier B.V. All rights reserved.
暂无评论