| As one of the two important aspects of satellite miss ion pla nning, satelliteground data trans miss ion is a NP- Hard optimizat ion proble m w ith mult iple time Windows, optimizat ion goals and resource constraints. Especia lly the us ing of the relay star provide possib le for all- weather and rea l- time data trans missio n, so, how to allocate the limited resources and vis ib le time windows beco me a problem demanding prompt solution. Three respects of work are done:Firstly, design Mult i- Agent system and establis h the satellite- ground data transmiss ion planning model. This paper introduces Agent syste m to the data transmiss ion tasks, design a federated and distr ibuted hybrid structure to realize mult isectoral task coordinatio n. And on this basis, build satellite- ground data transmiss ion pla nning model, descr ibed the basic assumpt ions, symbo l var iables, constraint cond it ions, taking the resource optimizat ion into account, so as to realize the collaborative plann ing of multi- type ground stations and relay stars.Second ly, design dyna mic planning pr inc iple and establis h the bidding mechanis m o f agreeme nt end ing contract net, For the dyna mic planning problems, this paper designs three kinds of scheduling princ iple, and introduces dyna mic integr it y to contract net protocol and improves tradit iona l contract net mechanis m to propose the agreement ending contract net. This paper furt her designs bidding mechanis m, contract net pro tocols, strategy of bidding and assessing bidding and al ocation algorithm ASD.Thirdly, present the improved Agent algorit hm and simulation experiments. This paper present the improved Agent algorit hm IAgent and deeply digging the relatio nship between the arriva l sequence and the p lanning perfor mance; the game re latio n between the cache and nested number. O n t he basis of t he simulat ion e nvironme nt, this paper processes compar ison ana lys is for mult ip le algorithms. |
PDF Full Text Request