Properties
For a CTMC and its embedded Markov chain , the classification of states is inherited from the embedded chain.
Let and "" denote the communication property.
Class Properties
| Property | Statement |
|---|---|
| Communication | in the CTMC in the embedded MC |
| State Partition | Since communication partitions the embedded MC into classes, it also partitions the associated CTMC |
| Irreducibility | The CTMC is irreducible the embedded MC is irreducible |
| Recurrence | State is recurrent in the CTMC is recurrent in the embedded MC |
| Transience | State is transient in the CTMC is transient in the embedded MC |
| Positive Recurrence | State is positive recurrent in the CTMC is positive recurrent in the embedded MC |
| Class Property | Transient and recurrent are class properties of a CTMC |
Interpretation
All structural properties (communication, recurrence, transience, irreducibility) of a CTMC are determined entirely by the behavior of the embedded DTMC. This is because these properties depend only on which states are visited and with what probability — not on how long is spent in each state.
Periodicity
Since in the embedded MC (the CTMC always changes state at jump times), periodicity is not possible in CTMCs. All states have period 1.
Related
Exercises
Back to Roadmap 📖 → 🃏 → ✏
Klasifikasi state dari embedded MC. Untuk , tentukan: (a) kelas komunikasi, (b) irreducible atau tidak, (c) recurrent atau transient.
Jawaban: Embedded MC: . (a) Satu kelas: (semua state berkomunikasi — 0↔1, 1↔2, jadi 0↔2 via 1). (b) Irreducible. (c) Semua recurrent (state space finite + irreducible → positive recurrent). Tidak ada periodisitas di CTMC.
Periodisitas. Mengapa CTMC tidak memiliki periodisitas?
Jawaban: Karena di embedded MC (selalu berubah state saat transisi), dan waktu holding bersifat eksponensial (kontinu). Transisi dapat terjadi kapan saja → periode selalu 1.