Causal computational complexity of distributed processes
File(s)DemangeonCausal.pdf (677.5 KB)
Accepted version
Author(s)
Demangeon, R
Yoshida, N
Type
Conference Paper
Abstract
his paper studies the complexity of
π
-calculus processes with res-
pect to the quantity of transitions caused by an incoming message.
First we propose a typing system for integrating Bellantoni and
Cook’s characterisation of polynomially-bound recursive functions
into Deng and Sangiorgi’s typing system for termination. We then
define computational complexity of distributed messages based on
Degano and Priami’s causal semantics, which identifies the depen-
dency between interleaved transitions. Next we apply a syntactic
flow analysis to typable processes to ensure the computational bound
of distributed messages. We prove that our analysis is
decidable
for
a given process;
sound
in the sense that it guarantees that the total
number of messages causally dependent of an input request received
from the outside is bounded by a polynomial of the content of this
request; and
complete
which means that each polynomial recursive
function can be computed by a typable process.
π
-calculus processes with res-
pect to the quantity of transitions caused by an incoming message.
First we propose a typing system for integrating Bellantoni and
Cook’s characterisation of polynomially-bound recursive functions
into Deng and Sangiorgi’s typing system for termination. We then
define computational complexity of distributed messages based on
Degano and Priami’s causal semantics, which identifies the depen-
dency between interleaved transitions. Next we apply a syntactic
flow analysis to typable processes to ensure the computational bound
of distributed messages. We prove that our analysis is
decidable
for
a given process;
sound
in the sense that it guarantees that the total
number of messages causally dependent of an input request received
from the outside is bounded by a polynomial of the content of this
request; and
complete
which means that each polynomial recursive
function can be computed by a typable process.
Date Issued
2018-07-09
Date Acceptance
2018-04-18
Citation
LICS '18 Proceedings of the 33rd Annual ACM/IEEE Symposium on Logic in Computer Science, 2018, pp.344-353
ISBN
9781450355834
Publisher
ACM / IEEE
Start Page
344
End Page
353
Journal / Book Title
LICS '18 Proceedings of the 33rd Annual ACM/IEEE Symposium on Logic in Computer Science
Copyright Statement
© 2018 ACM.
Sponsor
Engineering & Physical Science Research Council (E
Engineering & Physical Science Research Council (E
Engineering & Physical Science Research Council (EPSRC)
Engineering & Physical Science Research Council (EPSRC)
Engineering & Physical Science Research Council (E
Grant Number
ERI 025567 (EP/K034413/1)
PO 20015393
EP/K011715/1
EP/N027833/1
PO 20015391
Source
Thirty-Third Annual ACM/IEEE Symposium on Logic in Computer Science
Publication Status
Published
Start Date
2018-07-09
Finish Date
2018-07-12
Coverage Spatial
Oxford, UK
Date Publish Online
2018-07-09