Search citation statements
Paper Sections
Citation Types
Publication Types
Relationship
Authors
Journals
Heuristics are widely used for solving computational intractable synthesis problems. However, until now, there has been limited effort to systematically develop heuristics that can be applied to a variety of synthesis tasks. We focus on development of general optimization principles so that they can be applied to a wide range of synthesis problems. In particular, we propose a new way to realize the most constraining principle where at each step we gradually relax the constraints on the most constrained elements of the solution. This basic optimization mechanism is augmented with several new heuristic principles: minimal freedom reduction, negative thinking, calibration, simultaneous step consideration, and probabilistic modeling.We have successfully applied these optimization principles to a number of common behavioral synthesis tasks. Specifically, we demonstrate a systematic way to develop optimization algorithms for maximum independent set, time-constrained scheduling, and soft real-time system scheduling. The effectiveness of the approach and algorithms is validated on extensive real-life benchmarks. MOTIVATIONWe have two strategic objectives in this paper: development of general optimization principles and applications to a wide range of synthesis problems. The optimization goal is to advance the state of the art in the design of heuristic algorithms. Heuristics are widely used for solving computational intractable synthesis problems. However, until now, there has been limited effort to systematically develop heuristics that can be easily applied to a variety of synthesis tasks.In this paper, we propose a new heuristic optimization paradigm that can be applied on a broad spectrum of computationally intractable problems. While the traditional most constraining principle always addresses the most constrained part of the problem first, we employ the most constraining principle where at each step we make a decision that maximally relaxes the constraints on the most constrained elements of the solution. This basic optimization mechanism is augmented with several new heuristic insights: minimal freedom reduction, negative thinking, calibration, simultaneous step consideration, and probabilistic modeling. We call them the gradual relaxation techniques.The minimal freedom reduction principle aims to make the minimal possible quantum of decision at each step. The rationale * This research is partially supported by National Science Foundation under award CCR-0096383.is that after a small step is made, one can better evaluate its impacts and prevent the heuristic from following a greedy mode of optimization to produce local optimal solutions. The main way to realize this principle is to use negative thinking, i.e., to decide what the optimization process will not do at the next step, instead of what to do. The options that stay at the end of this process form a set of decisions to yield a high quality solution. Calibration is a step where the chances for optimization along a particular direction are evaluated ...
Heuristics are widely used for solving computational intractable synthesis problems. However, until now, there has been limited effort to systematically develop heuristics that can be applied to a variety of synthesis tasks. We focus on development of general optimization principles so that they can be applied to a wide range of synthesis problems. In particular, we propose a new way to realize the most constraining principle where at each step we gradually relax the constraints on the most constrained elements of the solution. This basic optimization mechanism is augmented with several new heuristic principles: minimal freedom reduction, negative thinking, calibration, simultaneous step consideration, and probabilistic modeling.We have successfully applied these optimization principles to a number of common behavioral synthesis tasks. Specifically, we demonstrate a systematic way to develop optimization algorithms for maximum independent set, time-constrained scheduling, and soft real-time system scheduling. The effectiveness of the approach and algorithms is validated on extensive real-life benchmarks. MOTIVATIONWe have two strategic objectives in this paper: development of general optimization principles and applications to a wide range of synthesis problems. The optimization goal is to advance the state of the art in the design of heuristic algorithms. Heuristics are widely used for solving computational intractable synthesis problems. However, until now, there has been limited effort to systematically develop heuristics that can be easily applied to a variety of synthesis tasks.In this paper, we propose a new heuristic optimization paradigm that can be applied on a broad spectrum of computationally intractable problems. While the traditional most constraining principle always addresses the most constrained part of the problem first, we employ the most constraining principle where at each step we make a decision that maximally relaxes the constraints on the most constrained elements of the solution. This basic optimization mechanism is augmented with several new heuristic insights: minimal freedom reduction, negative thinking, calibration, simultaneous step consideration, and probabilistic modeling. We call them the gradual relaxation techniques.The minimal freedom reduction principle aims to make the minimal possible quantum of decision at each step. The rationale * This research is partially supported by National Science Foundation under award CCR-0096383.is that after a small step is made, one can better evaluate its impacts and prevent the heuristic from following a greedy mode of optimization to produce local optimal solutions. The main way to realize this principle is to use negative thinking, i.e., to decide what the optimization process will not do at the next step, instead of what to do. The options that stay at the end of this process form a set of decisions to yield a high quality solution. Calibration is a step where the chances for optimization along a particular direction are evaluated ...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.