A SAMPLE-PATH APPROACH TO PALM PROBABILITIES

Previous papers have established sample-path versions of relations bet ween marginal time-stationary and event-stationary (Palm) state probab ilities for a process with an imbedded point process. This paper exten ds the use of sample-path analysis to provide relations between freque ncies for arbitrary (measurable) sets in function space, rather than j ust marginal (one-dimensional) frequencies. We define sample-path anal ogues of the time-stationary and event-stationary (Palm) probability m easures for a process with an imbedded point process, and then derive sample-path versions of the Palm transformation and inversion formulas .