How autonomy oriented computing (AOC) tackles a computationally hard optimization problem
08 May 2006-pp 646-653
TL;DR: An extended multiagent optimization system, called MAOSE, for supporting cooperative problem solving on a virtual landscape and achieving high-quality solution(s) by the self-organization of autonomous entities.
Abstract: The hard computational problems, such as the traveling salesman problem (TSP), are relevant to many tasks of practical interest, which normally can be well formalized but are difficult to solve. This paper presents an extended multiagent optimization system, called MAOSE, for supporting cooperative problem solving on a virtual landscape and achieving high-quality solution(s) by the self-organization of autonomous entities. The realization of an optimization algorithm then can be described in three parts: a) encode the representation of the problem, which provides the virtual landscape and possible auxiliary knowledge; b) construct the memory elements at the initialization stage; and c) design the generate-and-test behavior guided by the law of socially-biased individual learning, through tailoring to the domain structure. The implementation is demonstrated on the TSP in details. The extensive experimental results on real-world instances in TSPLIB show its efficiency as comparing to other algorithms.
Citations
More filters
TL;DR: Experimental results on small and large size TSP instances in TSPLIB (traveling salesman problem library) have shown that RMGA could almost get optimal tour every time in reasonable time and thus outperformed the known EAX-GA and LKH in the quality of solutions and the running time.
Abstract: TSP (traveling salesman problem) is one of the typical NP-hard problems in combinatorial optimization problem. An improved genetic algorithm with reinforcement mutation, named RMGA, was proposed to solve the TSP in this paper. The core of RMGA lies in the use of heterogeneous pairing selection instead of random pairing selection in EAX and the construction of reinforcement mutation operator, named RL-M, by modifying the Q-learning algorithm and applying it to those individual generated from modified EAX. The experimental results on small and large size TSP instances in TSPLIB (traveling salesman problem library) have shown that RMGA could almost get optimal tour every time in reasonable time and thus outperformed the known EAX-GA and LKH in the quality of solutions and the running time.
130 citations
01 Apr 2009
TL;DR: MAOS is refined for solving the traveling salesman problem (TSP), which is a classic hard computational problem and shows that the cooperative search of agents can achieve an overall good performance with a macro rule in the switch mode, which deploys certain alternate search rules with the offline performance in negative correlations.
Abstract: The multiagent optimization system (MAOS) is a nature-inspired method, which supports cooperative search by the self-organization of a group of compact agents situated in an environment with certain sharing public knowledge. Moreover, each agent in MAOS is an autonomous entity with personal declarative memory and behavioral components. In this paper, MAOS is refined for solving the traveling salesman problem (TSP), which is a classic hard computational problem. Based on a simplified MAOS version, in which each agent manipulates on extremely limited declarative knowledge, some simple and efficient components for solving TSP, including two improving heuristics based on a generalized edge assembly recombination, are implemented. Compared with metaheuristics in adaptive memory programming, MAOS is particularly suitable for supporting cooperative search. The experimental results on two TSP benchmark data sets show that MAOS is competitive as compared with some state-of-the-art algorithms, including the Lin-Kernighan-Helsgaun, IBGLK, PHGA, etc., although MAOS does not use any explicit local search during the runtime. The contributions of MAOS components are investigated. It indicates that certain clues can be positive for making suitable selections before time-consuming computation. More importantly, it shows that the cooperative search of agents can achieve an overall good performance with a macro rule in the switch mode, which deploys certain alternate search rules with the offline performance in negative correlations. Using simple alternate rules may prevent the high difficulty of seeking an omnipotent rule that is efficient for a large data set.
83 citations
Cites background or methods or result from "How autonomy oriented computing (AO..."
...Then, the quality can be measured by a simple R M rule [48], where ∀ xa, xb ∈ SR, if there is fC( xa) ≤ fC( xb), then xa has a better quality than xb, and R M returns TRUE; else, it returns FALSE....
[...]
...Due to its simplicity, the RT may determine nontrivial properties of the x (t) A [48]....
[...]
...The agents cooperatively search in an ENV [47], [48] for achieving a common intention of finding high-quality solution(s), based...
[...]
...the xO by a state-competing rule, which is simply realized by associating each candidate state xc with the competing function (fcomp) to be minimized [48], i....
[...]
...The IC is designed in a centralized way [48]....
[...]
25 May 2019
TL;DR: A multifaceted analysis of existing CSAS with learning capabilities reported in the literature is performed and a 3D framework is introduced that illustrates the learning aspects of CSAS considering the dimensions of autonomy, knowledge access, and behaviour, and facilitates the selection of learning techniques and models.
Abstract: Collective self-adaptive systems (CSAS) are distributed and interconnected systems composed of multiple agents that can perform complex tasks such as environmental data collection, search and rescue operations, and discovery of natural resources. By providing individual agents with learning capabilities, CSAS can cope with challenges related to distributed sensing and decision-making and operate in uncertain environments. This unique characteristic of CSAS enables the collective to exhibit robust behaviour while achieving system-wide and agent-specific goals. Although learning has been explored in many CSAS applications, selecting suitable learning models and techniques remains a significant challenge that is heavily influenced by expert knowledge. We address this gap by performing a multifaceted analysis of existing CSAS with learning capabilities reported in the literature. Based on this analysis, we introduce a 3D framework that illustrates the learning aspects of CSAS considering the dimensions of autonomy, knowledge access, and behaviour, and facilitates the selection of learning techniques and models. Finally, using example applications from this analysis, we derive open challenges and highlight the need for research on collaborative, resilient and privacy-aware mechanisms for CSAS.
41 citations
Cites background from "How autonomy oriented computing (AO..."
...agents actively collaborate to win a cooperative game [62], [63], smart sensors cooperate to patrol an area [37], or soccer...
[...]
...8 studies [29], [31], [35], [62], [65], [72], [75], [76]....
[...]
..., [62], [72], [75]) use domain knowledge for setting boundaries on what agents can learn at runtime and for accelerating learning, while in [28] the initial trigger...
[...]
...an environment holding a centralised socially-shared memory) which serves as a blackboard for all the agents (analogous to a shared past) [62]....
[...]
...Another subset of studies with similar learning task and associated emergent behaviour have system-wide collaborative tasks [27], [37], [62], [63]....
[...]
18 Oct 2008
TL;DR: The goal of this paper is to describe the key concepts in this computing paradigm, and furthermore, discuss some of the fundamental principles and mechanisms for obtaining self-organized computing solutions.
Abstract: Facing the increasing needs for large-scale, robust, adaptive, and distributed/decentralized computing capabilities from such fields as Web intelligence, scientific and social computing, Internet commerce, and pervasive computing, an unconventional bottom-up paradigm, based on the notions of Autonomy-Oriented Computing (AOC) and self-organization in open complex systems, offers new opportunities for developing promising architectures, methods, and technologies. The goal of this paper is to describe the key concepts in this computing paradigm, and furthermore, discuss some of the fundamental principles and mechanisms for obtaining self-organized computing solutions.
34 citations
TL;DR: This paper presents a new algorithm for the Symmetric TSP using Multiagent Reinforcement Learning (MARL) approach, which has a good performance with respect to the quality of the solution and the speed of computation.
Abstract: Travelling salesman problem (TSP) looks simple, however it is an important combinatorial problem Its computational intractability has attracted a number of heuristic approaches to generate satisfactory, if not optimal solutions In this paper, we present a new algorithm for the Symmetric TSP using Multiagent Reinforcement Learning (MARL) approach Each agent in the multiagent system is an autonomous entity with personal declarative memory and behavioral components which are used to tour construction and then constructed tour of each agent is improved by 2-opt local search heuristic as tour improvement heuristic in order to reach optimal or near-optimal solutions in a reasonable time The experiments in this paper are performed using the 29 datasets obtained from the TSPLIB Also, the experimental results of the proposed method are compared with some well-known methods in the field Our experimental results indicate that the proposed approach has a good performance with respect to the quality of the solution and the speed of computation
13 citations
References
More filters
01 Jan 1986
TL;DR: In this article, models of Human Nature and Casualty are used to model human nature and human health, and a set of self-regulatory mechanisms are proposed. But they do not consider the role of cognitive regulators.
Abstract: 1. Models of Human Nature and Casualty. 2. Observational Learning. 3. Enactive Learning. 4. Social Diffusion and Innovation. 5. Predictive Knowledge and Forethought. 6. Incentive Motivators. 7. Vicarious Motivators. 8. Self-Regulatory Mechanisms. 9. Self-Efficacy. 10. Cognitive Regulators. References. Index.
21,686 citations
01 Feb 1996
TL;DR: It is shown how the ant system (AS) can be applied to other optimization problems like the asymmetric traveling salesman, the quadratic assignment and the job-shop scheduling, and the salient characteristics-global data structure revision, distributed communication and probabilistic transitions of the AS.
Abstract: An analogy with the way ant colonies function has suggested the definition of a new computational paradigm, which we call ant system (AS). We propose it as a viable new approach to stochastic combinatorial optimization. The main characteristics of this model are positive feedback, distributed computation, and the use of a constructive greedy heuristic. Positive feedback accounts for rapid discovery of good solutions, distributed computation avoids premature convergence, and the greedy heuristic helps find acceptable solutions in the early stages of the search process. We apply the proposed methodology to the classical traveling salesman problem (TSP), and report simulation results. We also discuss parameter selection and the early setups of the model, and compare it with tabu search and simulated annealing using TSP. To demonstrate the robustness of the approach, we show how the ant system (AS) can be applied to other optimization problems like the asymmetric traveling salesman, the quadratic assignment and the job-shop scheduling. Finally we discuss the salient characteristics-global data structure revision, distributed communication and probabilistic transitions of the AS.
11,224 citations
"How autonomy oriented computing (AO..." refers background in this paper
...Each agent only has moderate problem solving capability by comparing with two extremes: a) the reflex agent, such as ant [12], which has no declarative memory and can only produce reflex behaviors in the environment; b) the cognitive architecture, such as ACT-R [1], which is rather sophisticated under unbounded rationality....
[...]
IBM1
TL;DR: A framework is developed to explore the connection between effective optimization algorithms and the problems they are solving and a number of "no free lunch" (NFL) theorems are presented which establish that for any algorithm, any elevated performance over one class of problems is offset by performance over another class.
Abstract: A framework is developed to explore the connection between effective optimization algorithms and the problems they are solving. A number of "no free lunch" (NFL) theorems are presented which establish that for any algorithm, any elevated performance over one class of problems is offset by performance over another class. These theorems result in a geometric interpretation of what it means for an algorithm to be well suited to an optimization problem. Applications of the NFL theorems to information-theoretic aspects of optimization and benchmark measures of performance are also presented. Other issues addressed include time-varying optimization problems and a priori "head-to-head" minimax distinctions between optimization algorithms, distinctions that result despite the NFL theorems' enforcing of a type of uniformity over all algorithms.
10,771 citations
"How autonomy oriented computing (AO..." refers background in this paper
...According to no free lunch theorem [50], the domain knowledge of the problem must be embedded at the implementation stage....
[...]
Book•
01 Jun 1972TL;DR: The aim of the book is to advance the understanding of how humans think by putting forth a theory of human problem solving, along with a body of empirical evidence that permits assessment of the theory.
Abstract: : The aim of the book is to advance the understanding of how humans think. It seeks to do so by putting forth a theory of human problem solving, along with a body of empirical evidence that permits assessment of the theory. (Author)
10,770 citations
"How autonomy oriented computing (AO..." refers background or methods in this paper
...As in the information process system [37], the declarative knowledge is represented in symbol structures, called T_INFO elements, and the procedural knowledge is represented in elementary information processes, called behavioral rules....
[...]
...INTRODUCTION The hard computational problems are ubiquitous in scientific and engineering fields [26], which often are conceptually simple and can be cast as a search through a space of alternatives [24][37]....
[...]
...The memory [19][37] is used for storing T_INFO elements....
[...]
TL;DR: This paper discusses a highly effective heuristic procedure for generating optimum and near-optimum solutions for the symmetric traveling-salesman problem based on a general approach to heuristics that is believed to have wide applicability in combinatorial optimization problems.
Abstract: This paper discusses a highly effective heuristic procedure for generating optimum and near-optimum solutions for the symmetric traveling-salesman problem. The procedure is based on a general approach to heuristics that is believed to have wide applicability in combinatorial optimization problems. The procedure produces optimum solutions for all problems tested, "classical" problems appearing in the literature, as well as randomly generated test problems, up to 110 cities. Run times grow approximately as n2; in absolute terms, a typical 100-city problem requires less than 25 seconds for one case GE635, and about three minutes to obtain the optimum with above 95 per cent confidence.
3,761 citations
"How autonomy oriented computing (AO..." refers background or methods in this paper
...The distance of two states is defined as the number of edges in which they differ [15][29][34]....
[...]
...The first is reduction [29], which locks the edges in the intersection of the edge sets (EI) of the tours for speeding up the subsequent search....
[...]
...To improve the chances for escaping from the local optimal tours, the double-bridge (DB) [29][32], a non-sequential kick-move, is normally used [2]....
[...]
...Although the TSP is easily formulated, it exhibits various aspects of hard computational problems and has often served as a touchstone for new problem solving methods [29][54]....
[...]
...2-Opt, 3-Opt and Lin-Kernighan (LK) algorithm [29] are representative LS methods....
[...]