D G - CAT E G O R I E S I V
by g a b r i e l d r u m m o n d - c o l e
transcribed by da mi en l ej ay
My topic is the proposition mentionned by Tae-Su: let
T
be a
dg-category and let
M
be a
C(k)
-model category such that
(1)
,
M
is
cofibrantly generated and
(2)
(the thing with tensor produ). Then
[T, Int(M)] is equivalent in Set with Iso(HoM
T
).
[T, Int(M)] ' Iso(HoM
T
).
You would like to do things with only dg-category
M
but unfortunatly
you need the homotopy tools contained in model categories. This
theorem is a theorem that links the dg-side with the model category
side.
On the left you have a morphism set and on the right you have a
set of objes.
Let me make a comparison with the following context: if
R
is a ring
and V a veor space, then
Hom(R, GL(V) ) ' {R module rutures on V}
It is not true that every dg-category can be written as
Int(M)
for
M
a model category. But we can always use the Yoneda embedding
described by Tae-Su.
h : T
0
−→ Int(T
0op
-Mod)
And the proposition applies to
Int(T
0op
-Mod)
. And now I can apply
the proposition to [T, T
0
] inside
[T, Int(T
0op
-Mod)].
We get that
[T, T
0
]
can be identified with ‘those modules’ whose image
lies in the quasi-essential image of h.
Let us see an example: take
T
be the dg-category with only one
obje and Hom
T
(∗, ∗) ' k. Then
[k, T
0
] ⊂ Iso Ho(ob of T
0op
-mod)
And those would be the isomorphism classes of objes in
T
0
i.e objes
of [T
0
].
The Derived Seminar, June , Pohang, Korea.
©
Damien Lejay. All rights
reserved.