摘要:In this paper, we proposed a mechanism named integration of scheduling algorithms and built architecture of process-type MASS to support it. The mechanism is system-based. It can take into account of the complexity of real-life systems and integrate almost all of scheduling algorithms into one system architecture and then use them dynamically to adapt to the change of production environments.
d highlighting future extensions.
2. INTEGRATION OF SCHEDULING ALGORITHMS
Different scheduling algorithms have their own feature and using condition or area. As units, their capabilities are limited. But they can be used to solve complex scheduling problem if only they are combined together. Integration of Scheduling Algorithms (ISA) is such a process or mechanism of integrating various scheduling algorithms to generate a single architecture, which process more applicability.
To some extent, some mixed algorithms, which are compose of other algorithms, are particular instances of ISA. But these combination are simple and immovable, and the result, which are also single algorithm, is still limited to some special production environments and we are not interested in it here.
ISA that be defined here is a flexible and intelligent scenario of collaboration among scheduling algorithms. It can dynamically call or joint different algorithms together for different environment under various conditions and constraints. Each algorithm (we call it atomic algorithm) joining the scenario is independent.
There are two cases of ISA. One is that atomic algorithms store in a system are dynamically called to response the change of environments. The other is that scheduling problem is divided into several subproblems. Each subproblem is sloved by one atomic algorithm. The whole problem is solved by dynamic collaboration of several atomic algorithms.
We may design an intelligent system that consists of a rule base, an algorithm base, a reason machine and other components. The rule base stores knowledge of using algorithm. The algorithm base stores all sorts of scheduling algorithms. The system can make estimation in reason machine and select a good algorithm from algorithm base according to rules from rule base. Theoretically it can solve any complex scheduling problem. But its efficiency is hard to ensure because we have too many algorithms and rules of how to use these algorithms. To retrieve rules and algorithms need very long time.
Then, of course, we think of distributed structure. MAS is surely a good architecture to help realizing ISA.
3. MULTI-AGENT BASED ISA
A scheduling algorithm is a process of solving scheduling problems. The process needs to keep contact with the environment. Assembled with a rule base, an analysis module, an interface (to contact with outside), and a reason machine, etc., a scheduling algorithm can be characterized as an intelligent agent. The agent can make decisions based on the response from the environment and take action (computation). We name this agent computing agent. The dynamic integration of scheduling algorithms is the integration of different computing agents under the scheduling of a manager.
3.1 DEFINITION OF AGENTS
A whole computation consists of several steps: environment analysis, goal setting, evaluation of computation capability, decision making, computing, output conclusion, etc. We integrated these steps into a general model of computing agent as Fig.1
Elements of a computing agent are detailed as follow:
1) Algorithm Base
Stores algorithms that belong to the same
本论文由英语论文网提供整理,提供论文代写,英语论文代写,代写论文,代写英语论文,代写留学生论文,代写英文论文,留学生论文代写相关核心关键词搜索。